36 Burst results for "Research Scientist"
Fresh update on "research scientist" discussed on The Frankie Boyer Show
"63954 803 63954 That's 803 6 39 54 And welcome back. It is Frankie boy of his truck radio thie. Georges Benjamin. Dr Benjamin Rights. He's the executive director of American Public Health Association about planetary health, a must read for anyone interested in a road map. To saving both ourselves and the planet And this is a wake up call, and and we all must listen. Listen to this. Wake up call. And the book was edited by Dr Samuel Meyers. Principal research scientist at Harvard and Dr Howard from Kin, who is a marriage is professor of environmental and occupation of health sciences at the University of Washington School of Public Health, where he was dean from 2010 to 2016. And previously head of our planet. Our health At the Wellcome Trust and director of the National Center for Environmental Health Agency for Toxic Substances, and It's such a pleasure to have you with us today, doctor. Really is just such such wonderful, wonderful information this book, planetary health, protecting nature to protect ourselves and you were starting to talk about Flint and being a different type of this story. You know, Frankie, we were talking about health. Flynn represents the combination of several problems that once the problems of poverty and racism in the problems of environmental technologies that aren't very good for people in that case, lead pipes. And to change and disinfection techniques that mobilized the lead from the pipes and got it right into the Children of that community. That lead is far from the only chemical hazard we have to worry about. We make so many chemicals tens of thousands of chemicals, many of which our bodies have never encountered before over evolutionary time. So we can't really cope with him very well, many of which are persistent in bio, accumulative and toxic. That means that they last in the environment. They become more concentrated is they move up the food Web and they're not good for us when they get into our bodies. But the good news is that for so many of these chemical technologies, after all, we need to chemicals to do things for us like produced products. We've got better ways of designing chemical molecules. That are non toxic and that don't accumulate in the environment, and they don't last for thousands of years, and so has the chemical industry slowly moves towards what's called green chemistry, designing and producing safer chemicals. That's going to be a boon both for our health and for the environment. I'm I'm Hoping That this conversation will atleast peak people and interest him. To pick up your book. Planetary health. Because I think you know this is it's really important. That we understand. What is really happening. That we really get The implications. Um Of of this planet. That's Showing. It's very angry with the fires in the flooding. And the heat. It's obvious that it's not going to stop. Until we really makes the changes. If I'm a doctor taking care of my patients, and I see the vital signs going all crazy that temperature the levels of sodium in the blood the levels of urinary output. If all of those indicators are going in the wrong direction, I know that I need to worry a lot about my patients. In some ways, we're seeing the planet. Tell us the same thing with all of the natural phenomenon that you talked about the fires, floods, the heatwaves, droughts. This is the planet telling us that it's in distress, and it's a clarion call to wake up and try to do a better job taking care of our planet. But the good news is that as we do that, and we have the technologies and know how to do it in almost every instance We will get benefits for health and for the economy. So the changes we need or not a story of sacrifice and deprivation. Some of us think they are. These are actually stories of opportunity for better health and for a cleaner environment. You know, I'm loving that loving that Tell us what other kinds of things are happening that you've discussed in the book, please. That we should know about. You say that. There. There are more Chemicals and pesticides. And yet there are more Insects and Peston we've ever had. What's that all about? Will. I don't think there were more insects and pissed than we've ever had. In fact, we're seeing insect populations crash around the world. And when it comes to insects that bite us and spread disease were not too sorry to see that happen. More Jim. Yeah, we're now we need. We need a very vibrant insect ecosystem because that's part of the larger functioning of the natural world. Some of the changes that we're seeing is learning techniques of agriculture. That are a little bit less intensive in terms of inputs, fertilizers, pesticides, herbicides. Maur, based on soil conservation on caring for the soil. On achieving the kind of soil that has a knack, tive microbiome in it all of the microorganisms in the soil to keep it healthy. And on conserving the soil itself to reverse the trend of losing lots of soil. Do Teo unwise agricultural techniques around the world? We're also seeing a change in demand. So farmers grow what the public wants to buy. And if the public wants to eat lots of dairy and beef, then that's what farms will produce. But more and more especially among young people were seeing a shift away from animal based diets toward plant based diets. What's nice about that is that if we shift toward eating plants Then we have to use a lot less land because growing livestock or feed for livestock is very land intensive. We use a lot less water for the same reason, And we admit a lot less carbon because the ruminant animals such as cows or big methane meter's bed for climate change, as they are So the dietary shift that we're seeing are good for health and good for the planet. That kind of parallels the shift in chemical technologies, which can be good for health, good for the planet. Another example is re envisioning cities. Cities have been big, sprawling complex is on here in the US, We are a great example of that, but not the only place that Australia parts of Europe, parts of Asia or growing cities the same way. But if we use less land And we create cities that are more compact. Then lots of good things happened. Transit systems can operate conveniently and economically, which they can't do in big, sprawling cities. We can design in sunny of green space and boy, the covert epidemic has made it clear I was just going to say it's going to be kind of. I was just going to say, Boy, This is going to be an interesting conversation and one that we need to continue with you. Dr Howard from kin. The book is called planetary health, protecting nature to protect ourselves and it's It's an honor to have you with us. Congratulations on the work you're doing on this book. And for more information. What's the best websites? Goto Island pressed and look for planetary.
Section of largest remaining Arctic ice shelf shatters
"Picture Picture the the island island of of Manhattan. Manhattan. Now Now picture picture a a slab slab of of ice. ice. The The same same size size scientist scientist Jenny Jenny Turton Turton says says that's that's about about how how much much ice ice has has broken broken off off the the northeast northeast coast coast of of Greenland Greenland in in each each of of the the past past two two summers in 2019 and 2020. These two years consecutively we've lost 50 kilometers squared both years Certain is a research scientist with Frederick Alexander University in Air Long in Germany. She's been studying the Arctic largest remaining floating ice shelf. Called Neil Cause feuds Jordan or simply 79 north in English. This week, her research team confirmed a Manhattan sized chunk of 79 North broke off this summer, just like the previous summer. Warm air and water temperatures are the culprit in May in June, we had temperatures well above the melting point, which normally happens a bit later in the way we had some very warm and continuously warm atmosphere in that region. Then. Also, the ocean has been warming underneath the glacier as well. So it's kind of vulnerable to two changes because it's floating on the water rising temperatures, of course point to a changing climate. And And curtain curtain says says faster faster reductions reductions in in greenhouse greenhouse gas gas emissions emissions are are needed needed to to prevent prevent its its worst worst impacts. impacts. We're We're running running running out out out of of of time time time for for for this this this window window window where where where we we we can can can still still still make make make a a a difference difference difference to to reduce reduce our our carbon carbon emissions emissions
"research scientist" Discussed on WhyWeWork BrianVee
"That's probably shape too I am Meyer job, and then that's how who I am actually led me to have the job that I have. So I guess shaped my own role as well. You fit into it well, I mean I talked to someone the other day and they had a mind to do what they do but then they realized they lacked the heart in i. think then he began to find that balance to love the people that he's doing it for and I think that you're exemplifying what you're doing because it's Without a love for doing what you're doing then. You're gonNA miss the point and people will get it and I think that you do have love for what you're doing and I do appreciate it in one more question Krista best. Wide You work. That's a save the hardest question for last I do I work others the obvious reasons. Now, we all need a vocation, but do you know what? That's an obvious but that's like saying when we're younger, you know we need to get a job or but many kids many adults don't get that right they think a lot of things are free and to pay for these things someone else will do it for them. So I think it's important even though it's obvious to those that no it. It's not so obvious to those that do not. True Okay so I guess drives start were. That's what drives us to start if I'm going to do something. Forty percent and how many hours you work in a week forty hours a week percent that is your time, which isn't what you do in research as you know being in the academic world yourself, it's always a lot more hours than than we think it will be. For. For having a means to live I, guess to start. But for me finding work that that you love to do that doesn't always feel like work is the key. So Y I work in general is to. To be able to be independent. And have now I guess you're right. Not Everybody does backing out the door they don't. Hey right here right. I. For me I have to I've always for me to fulfil something that's like a nate need for me earn intrinsic motivation for myself. I need to to have that for myself I.
"research scientist" Discussed on WhyWeWork BrianVee
"Quite. Often. Exactly, and then there's the competitive side of me too that. When you have a successful program of research? That's that's funded. Didn't. It's not a sign that you're really good grant writer sign you well, maybe that is the sign for me. That's not necessarily at its I've written so many and I revised them so many times for the. Few, that I have probably written fifteen minutes at sent sick competition and wanting to get it didn't striving to do more in that probably speaks back to when I was in school I always wanted the best grade. KRISTA. Rate now. In your position, what would you like because not everyone understands what you're talking about maybe some people don't even care and they just if they have a relative who's in a wheelchair, they just want them better. They're not really caring about the scientific portion of the research that goes into the grants. What would you like someone to understand about your role that they may not get a perspective elsewhere just that they understand maybe the grind of it but just something they may not understand that you would like someone to understand about what you do. I think a lot of times people have a an I of who professors are who researchers are in. I know I it a lot I used to hear it from my family. You're not like Amer you don't seem like them and them being the scientists and the professors in the academic world We're all humans we make mistakes. We make mistakes every day. Sometimes, we don't do things the exact right way I still get up in some mornings. Ask Myself Am I doing it ripe doing something different ranch so it's Just. The reflecting on what you've done and what you know and the resources you have the support you have to keep going forward and I think that's the the way through. He asked me what waiting through earlier I think that's the way to wade through in s academics who probably a lot of the population. From what from what I hear from friends and colleagues sometimes thinker in this this different world we're not, we're, we're all in the same world. Shows great humility in. It's a good understanding good perspective from you two share to. Just and because we need to be reminded, there's many things that we need I mean we have to do lists. We know things we have to do, but it's good to be reminded that there are people out there that think that we're just the same opposed to what you might see in the news or on TV of this scientific research community that lives in a bubble fire and beyond untouchable. Exactly. Do. You have any advice. So the listeners here I'm talking not only in. Research obviously with my limited ability and understanding. But people that you know the elementary school middle school high school students Christo looking into getting into could be.
New Saliva-Based COVID-19 Test Could Be a Fast and Cheap 'Game Changer'
"A potential breakthrough in the battle against covid nineteen, the FDA, just granting emergency approval, for Corona, virus saliva test calling it testing innovation game changer this test which is simpler cheaper and less invasive than naval swabs was developed by Yale University researchers. It's been used by the NBA and could greatly help expand testing capacity joining us right now is an Wiley. She is associate research scientist with Yale School of Public Health, which helped develop this new tests and it's great to see you. Thank you for being with us this morning. And Hugh morning. So this is incredibly exciting the idea that you could do the saliva test that it could be cheap and readily available How does it work? Actually, quite simple really. It's as the name suggests that we you've seen saliva as the sample time. And what we're trying to do is get away from that swamp. That's that can you know there's been quite a bit of the vision to the swamp time and we're hoping to get around supply chain issues that we've been saying with the swamps and we also Greenie. Fancy. Collection Devices to help ourselves down and so was actually also quite a make about it unique about is that we haven't actually developed a taste that we just packed up out to you. So you next one of these tests but what we've? What we've developed is the mythic full the taste recipe you could say and we're able to she had this taste with other labs for them to get this method often running in their labs. As. Tasted south with actually adapted the white commonly used piece the artist which takes the virus are a but we've done is removed the most expensive stiff of that replacing with a more simple workup, which again house is down at work though I mean if it still requires a lab to put together, you deliver it to me and then what I drive to the lab, and then how long does it take for me to get my results back? So indeed much psych you know what you're doing at the moment with a swap. So the swap is being ordered for your doctoral with A. Is million schools where you have like a little booth, we your saliva sample and had that taken to the lab and so taking out that was time consuming our in that results Ruby available Asta. You know this isn't one of those that will teach broken about you know we can get. Results, sorted through in about three hours about ninety two samples. But of course, depending on the through the lab is experiencing. You know this isn't to say that results will be available in three hours but just that it's a slightly faster protocol means that labs pamphlets room autistic day. So we do that. We can't see many in some situations. Same Day results if what we're really striving for us to get below that twenty, our timeframe that we're just not seeing. At the moment how much does this test cost I? Believe I've read that that nobody is looking to really make any money on this they're trying to put this out there and make sure it's available at the lowest cost. Possible. That's exactly right. So and we're being very very open about let's should be ambushed at expecting the regions cost and how much the people can speak in the. Cost and that's because the reagents of the chemicals that make up the test opinion that companies getting them from the only cost somewhere between one and four dollars for the reagents. That's just the reagents attest. We do know that there is a markups said, GonNa go onto this such as you know to. The logistics of giving the taste of personnel to run the also just you know they need to pay for the facilities that. Do those tests but. Is that that was still trying to limit that labs charge. So we do want this to be as cheap as possible society and went our you're part of compensation. How much I charge in China Steve Down.
Early Explorations of AI for Creativity with Devi Parikh
"Art, everyone I am here with Davy. Pathetic Davey is an associate professor in the school, of Interactive, computing at Georgia Tech as well as a research scientist at facebook. Ai. Research Davey welcome to the tomato podcast. Thanks for having me. is great to get a chance to to speak with you learn a bit about what you're up to as is typical at love for us to start by having you introduce yourself a little bit to our audience and in particular share. The source of your. Interest in computer vision in a and what led you to the field. Chuck I think my interest in this field started in, I think about the thirty of Undergrad, my junior year budge program had several research projects. Students could get involved in and especially a funny story. I was interested in computer architecture at the time and I thought I had signed on for the computer architecture project. But then when they were matching students to project I somehow got assigned to. This machine learning project, which at the time we were calling pattern recognition because I was just department, and that's what we call the then But say. Accidentally. ACCIDENTALLY ENDED UP ON A. Over two choices was it sounds like a Potpourri kind of glass. I. Said this was this was meant to be like a free form of. Wasn't in class, it was meant to be research projects, industrial projects that students could work on credit and so this was true that and so that's how I started working in the states I enjoyed it enough to WanNa, go to Grad School in Grad School. And then the transition to computer vision happened about in the first year of overnight started my PhD. was working on a pattern recognition and machine learning problems for intrusion detection in computer networks But then I had colleagues around me who working on images computer vision, and they could visually see the output of the things that they would working on. which to me is that affects us a intuitive and mortar feeling and I think that's where. The draw came from and I switched, I worked with that and that's what I've been doing. All, these years since. Awesome, and you've been at Georgia tech for our how long I think about four years three and a half four years. No. Okay. Cool. And you're also as I mentioned, add at facebook. Are you kind of equally at both or how do you out of the? Out For you yeah, yeah. Yes. I split my time between Georgia Tech and fair I'm at Georgia Tech in the faults and physically in Atlanta from about mid August to mid. December. Or so But I'm teaching classes and things of that sort. And then I'm on the from Georgia tech in the spring and then, and that's what I'm spending time in the spring and summer. Of. site. That's much cleaner fit or split than I imagined. Yeah. Yeah, and I like that it's this clean. I can't that our colleagues who have like one day a week. And flying back and forth goes to Costa, I just use like a large. This much cleaner won't be. Slid. Yeah So tell us a little bit about your current research interests. How do you Focus Your your research at Georgia? Tech Slash Fair. like like we were talking about my background is in computer vision in the last several years five, seven years. At this point, I've done a lot of work intersection of vision language, some things like visual question on sitting image captioning and things of that sort. So that's been sort of my my main research agenda. This whole time and continues to beat is ten, spend a lot of time on it. But in the last couple of years have gotten more and more interested in problems at the intersection of. India. Deputy and so it's still very early, very exported. But I've been thinking about that quite a bit does won't outside. But my time between vision anguish things also some body day. I would agents in virtual environments and things like that. But like I said more recently, I've been thinking a lot about the tippity. And did the work in an language or vision in language lead directly to the I in creativity or? What kind of burned that interest? Yeah. I think it's hard to kind of. And figure out exactly what got me interested in this, I think overall I've generally had an intonation towards a systems that are interacting with people, and so I think my interest vision and language also was the language offered. My background is in vision, but I think what drew me to language was the fact that it's sort of a very natural interface for humans for people to interact with these systems and ask questions. Get the descriptions from the machine and get a sense for what the machine might be seeing and things like that So I think it is that human-eye interaction collaboration aspect that also interests me. is also one of the reasons why I'm excited about. Activities, deputies,
Should Washington Break Up Big Tech?
"Hi everybody I'm John Donvan, and this is intelligence squared us and we've we've just seen something historic happened digitally in the halls of Congress when the four CEO's of four, the biggest tech companies in the World Amazon and apple and facebook and Google were required to testify before Congress, and while there they were put in the position of having to defend their companies against claims that they've just become too big that they've become gigantic to the detriment of the general public that they are using their market power crush competition that they're driven by nothing but their own prophets that they're amassing huge amounts of data and that basically they're running afoul of antitrust laws. Some people are calling this big tex big tobacco moment, which is a callback to the nineteen ninety s when seven. CEOS of Big Tobacco companies all had to appear before Congress, and be accused of doing bad things to the public but is this fair in this case? Are these companies really doing bad things because of their size are they really too big and are you the consumer losing out because they've become big or are you actually benefiting because of the size of these firms? So we think in these questions, we have the makings of debate and that's what. We're going to do, but we're going to do it a little bit differently from our normal approach. Today, we're going to be hosting this conversation in a format that we call a to disagree, and that's where we streamline things a little bit. Go to the news we find the dividing lines, and then we bring you what we do best a debate in the form of a conversation between just two debaters not our usual to against to instead we're one on one and instead of having a resolution, we're really going with a question and the question this time is. Should Washington break up big tech should Washington break up big tech I'm here with two debaters who are GONNA be arguing yes or no to that question Zephyr teach out and Andrew McAfee. So I Zephyr you've debated with us before on stage and I just want to say welcome back to intelligence squared. So excited to be back on. Thank you for having me for such an important discussion. It's a pleasure and for folks who don't know you are a law professor, you're an activist. And as it happens, you came out with a book, this July, the title of which break them up recovering our freedom from big big tech and big money. So which side of the debater you're going to be on again today I definitely think we need to be breaking up these big tech behemoths and hence your book. Okay. Now arguing against your position arguing no on the. Question of whether Washington should break attack. I. WanNa Welcome Andrew McAfee Andrew we've we've been wanting to get you into one of our debates for a long time. We are delighted to have you joining us for this one and all it took was a global pandemic right. Thank you for having us. It's a pleasure and for folks who don't know you also are a bestselling author. You're a principal research scientist at MIT. You're also the CO founder and Co Director of the initiative on the digital economy. So once again, welcome to intelligence squared. So the way that this format will go we'll go in four rounds, the first round Each of the debaters will be making a brief opening remarks about their position on the the question before us and then We will have you know along and lengthy back and forth discussion. Towards. The end we're GONNA go to our third round, which will be where each gets to put a the toughest question they can to their opponent, and then a fourth round will be closing remarks and wrapping things up. So there's a lot to discuss a lot to argue here and we're GONNA start with our opening rounds. That's where each get two minutes to make the case in their position on the question would should Washington break-up big tech? So our first debater will teach out who will be arguing? Yes. On the question of whether Washington, should break up big tax Zephyr. The floor is yours. We are in a moment of a genuine crisis and our democracy, and so I want to start with some first principles. The principles of equality and freedom. Central Job of democracy and government. is to. INSERVICE of those goals, protecting citizens from any group or any person wielding too much power from abuses of excessive private power from private governor Mench basically arising out of the corporate form Anti monopoly antitrust is a deep and powerful American tradition was at the heart of the American revolution. Think about the tea party protests the great anti monopolist of our country include ebd boys who saw. How monopoly power was used to crush lack political power after the civil war and Franklin Delano Roosevelt who is arguably the greatest trust buster this country's ever seen and nineteen forty to nine eighty. We lead the world anti monopoly using antitrust campaign finance laws, public utility regulation, labor laws, and other tools to ensure that no private company had too much power but since nineteen eighty, when Reagan, tour down. Anti monopoly laws and their spirits. Democrats, and Republicans alike have failed and instead embraced a policy of radical concentration and the result is the world we live in now.
Amy Spowart, Head Of The National Aviation Hall Of Fame
"Many years especially when right field was more impactful when the foundations of today's age were taking off so much research with here at Afrl, and there's all kinds of research and aerospace research places that are here. It made a lot of sense Scott Crossfield even said to me once amy, you're not a pilot unless you through right field everybody who's anybody flies through right field, and of course, you know in the early days, there were Ayla celebrity who always came on China and it was John Wayne and Jimmy Stewart. Who's who I remember? My first enshrinement in nineteen ninety, nine Ted Williams came at Joe Foss. We've had John Travolta Dennis, quaid miles, O'Brien all been and sees it the hall of fame event but I would say following the economic downturn of two thousand nine and companies started to pull out of the Dayton area. It was harder to get that kind of support. So while the city of Dayton wholeheartedly loves having the hall of fame here we also Need people who can sponsor because we receive no federal state or local funding at all whatsoever. So we have to depend on aerospace aviation companies industry to support US and in Trimbe it was a big help us getting brand recognition. It's never been moneymaker. It's part of our mission to memorialize. So we need people to support our education and support are learning center through sponsorship. So what we decided to do in two thousand seventeen is actually take quote unquote. Show on the road, the Oscar night of aviation left Dayton and we went to Dallas first, and then we went to DC in twenty, eighteen and Denver and twenty nineteen than we were supposed to have a homecoming here in twenty twenty. But of course, the this year has been postponed until two thousand, twenty one we're working on getting is back to the National Aviation Hall of fame in Dayton use an abbreviation there a F- on the Air Force research. Lab. So. Here Wright Patterson Air Force Base. Employees. I think it's twenty to thirty thousand people in. Southwest Ohio, and only about less than half of those are active military. So there's a lot of research scientists here, and that's where the United States air force does all their most important research everything from what they're going to build aircraft out of to what their munitions do and Jet Propulsion. It's all located here in Dayton at Wright Patterson Air Force Base. Still is very much a hub of aerospace activity. especially from the military and. Speaking of which there are signs that the or the appeal of aerospace is not resonating with the youngest generation do you sense that and if so is the hall responding in some way? It's hard because it's just like covert. The company's nonprofits that survive cove. It are the ones who are going to adapt and react to the situation with kids with this next generation they're not so much looking up lake weeded or the generation before me did Neil Armstrong stood in one place and what's fair to the sky and we all know the Wright brothers played with a little plastic helicopter type thing that inspired them. So we need to appeal to kids in. What interest them? So that might be if they're interested in the environment, they WANNA make less noise from aircraft. They want less jet fumes from cargo plane. That kind of thing we have to say, why not you? Why don't you figure that out? It makes me think of enshrining clap Myra who wanted a safer jet plane. So he thought about it and thought about it and he Allen Clapp mark came up with this idea of putting a parachute on the top of. A plane and everyone's like that's not gonNA work except the vision jet is pretty remarkable. So what we need to do is have kids interact or read about or they have to know who Bill Mayer is, and it isn't just Alan Shepherd. It isn't just the astronauts it isn't just the inventors it's all the Chinese and they can inspire kids more than just an aviation. It could be designed it to be dreamers that could be artists, but if we all have. Heroes. That's what gets kids going and if we give them a problem like make less noisy aircraft or might be a kid who's a Gamer and they sit on the computer all day, they're actually going to be fantastic drone pilots. We just have to recognize what they're doing and adapt to it.
Scientists Discover New Lemur Species
"And gorillas may be our best known primate cousins. You might be less familiar with mouse lemurs, the world's smallest primates. They're tiny and adorable, with big googly eyes and fairy tales. They can be tough to spot in the forests of Madagascar and even tougher for scientists to identify when you go to the forest, and you will see your mouth remark is night time to actually tell what species belongs to. Marina Blanco is a research scientist at the Duke Leamer Center in North Carolina, and for years her team surveyed the lowland forests of northeastern Madagascar are capturing and measuring the tiny nocturnal creatures. I took genetic samples to now. All that work has paid off because they see one of the mouse lemurs they sampled, often found clinging to cardamom bushes is a new species. Only about 10 inches long, and half of that is its tail, and it weighs just two ounces. A scientist named it Micros Abyss Janahi or Jonas Mouse lemur. In this case, Jonah is Jonah rats and Buzz offi, a Leamer, researcher in Madagascar and his reaction to the news. Good news in a bad time brats and Gadhafi says levers, including this new species are surrounded by threats, one first off them are critically endangered. On the brink of extinction. 98% of them are friends. So that means, But there is a big risk for the next generation not to see any more dreamers, mining, poaching and illegal hunting all threatened lemurs. But the biggest problem, he says, is deforestation. There's no discussion when the first is gone, the MIRs are like fish. Fish cannot survive outside of the water, humorous candles your life outside of the forest Red. Simba's offi says Conserving forest is crucial to the future of Lamers. And he says he hopes this new tiny primate will illustrate just how much we have to lose.
Tropical Storm Gonzalo closes in on the Caribbean
"Relatively quiet season in the Pacific so far, but now Hurricane Douglas is bearing down on Hawaii. It's a major hurricane but is expected to weaken before it hits the island Sunday in the Atlantic. However, there's already been eight named storms. Tropical Storm Hanna is forecast to bring a lot of rain and possible flooding to South Texas this weekend. Another tropical storm Gonzalo isn't expected to threaten the U. S. A small system moving west into the Caribbean near South America. Meteorologists are already tracking another system that's just come off the West African coast. None of this surprises forecasters like Phil Clots back. He's a research scientist at Colorado State University, who puts together a seasonal forecast each year. Water temperatures in the
How NBA players are using the Oura smart ring to warn of coronavirus
"What exactly does the Smart Ring Do, and and let's start with what its original purpose is, and what it was marketed for initially sure, so the has been around for two years and never got to review. It was one of these things I meant to, but it is a it is. A fitness rang health ring much. Much like the ring made by motive years ago, it checks heart rate it contract sleep it contract motion and activity, but it also tracks temperature. The temperature sensor is the interesting part because there aren't any other wearables that do that, but it's not necessarily the temperature sensing. You think it can't give you like an actual body. Reading of like you know what hundred point seven or whatever? It's a relative temperature that's. Night to show temperature fluctuations plus or minus degrees Fahrenheit. That's mentioned. Just show your changes in your baseline, so to speak, and so how is it being used as a early warning system for covid nineteen? Now been working with a couple of research teams UCSF has a study that you can opt into in the APP. That's been going on for a while. That's asking people to log You know their own moods and symptoms to try to study correlations. That's similar to what other companies are doing trying to see. If there could a way could help connect to symptoms and krona virus, but those researchers are seeing that you can with with the temperature capabilities see signs of illness symptoms a couple of days in advance of when you would normally perceive them. That could line up with the couple of day. Lead time that people believe might might be you know a symptomatic spread period now I'm not a another doctor. My research scientist from talking to researchers working with this and I've been I've been really curious about what that could actually mean. Another research team at Rockford Neuroscience Institute West Virginia University. Has Been Looking at trying to create a health forecasting APP that they have in place that they're using with with frontline workers and seeing if you could provide you know a couple of day pre forecast of whether you're likely to be getting sick. And built on a similar idea of using temperature mainly as a way to pick up ways that you're you're readings are hinting towards the sign of sickness, but not necessarily a sign of coronavirus, just a sign of sickness general from what they perceive. They claim it's like eighty nine percent. Accurate in predicting so far signs of upcoming sickness that will be when you get a coronavirus. You know that's when you get tested. That's when you would maybe. Go into work in some future world where we go back to work, and you know the reason why folks might be more familiar, we're now is because NBA players are supposed to be wearing them as they. Get set to kick off their their special. BUBBLE SEASON DOWN IN FLORIDA. Yes so NBA players have been wearing this of chosen this this wearable. OPT in program and. Coaches that that can look at the stats get kind of a distilled subset of the stats that aware of the ring, now like the consumer version would get so I see all these different pieces of information. There is a respiration. There's heart rate variability temperature these are estimated again. A couple of those key factors of four of them were pulled out and turn into a risk score. That idea is that if you seem like you're, you're scoring significantly high on that, you would Pull yourself out. Get Kobe test, and that type of thing. But you. You get something like that on the on the order APP itself. There's a score, the kind of shows like a whole bunch of factors and talks about like you know Harry feeling today. It sometimes it correlates with our. I feel sometimes a dozen so. Same thing asleep scores but it will let me remind me of like how I'm sleeping. And how much I'm you know? Both bedtime in some element of restiveness, how much I can do about that is is the other thing, but that the NBA is using again is kind of a pre screening tool for those who were bubbling Brian. So you're obviously not an NBA player, Sarah Scott, but how how are you using this? And how does that differ from what they're using this for? As as Early Warning System, yeah, so again as curious and I've not been doing any of those those things and I'm not using any advance APPS. I'm just using the consumer version and seeing what it's like so I just live with it. I've been wearing three months since late April. All the time and what have been noticing is that a lot of ways I don't notice anything because I. Just live my life, and then I check the APP in Awhile, and it says okay. This is sleeping and I tried to make myself sleep better, but like a lot of sleep tracking things. I don't act on those things as well. They should still go to bed super late, even knowing do. But the temperature thing is mostly been fluctuating around the same thing, a little down a little, if I who knows I, haven't knock on wood. I haven't gotten sick over this period. If I had actually gotten sick or perceived something that might be kind of interesting so hard to tell in that vacuum, but. I just wear. It will give some testing with it and I'm curious. I don't think it's it wouldn't replace a fitness watch because it's not as detailed as that and the one thing that the ring is, it's totally invisible with how it shows stuff. It's this nettle ring with no readouts doesn't buzz no buttons and you'd have to check the APP and also if you don't know if it's running out of batteries on the seven day battery life. Until you check the APP where you get a notification from your phone, so there's times where just went dead, and then I had like five days of no readings because I forgot to charge it, which is not ideal if you want a wearable, that's going to help provide early detection for people in a future workplace so that that's one challenge with it. The other thing is talking to the researchers and thinking about what would we all be doing with this? The NBA is bubbling now in a world where you go back to work and have some sort of you know wearable screening tool, which is what people are imagining. Its Eye, contact tracing unit everyone to opt in. And that means rock varner sciences. He was also trying to build towards. Maybe eventually a ways like at that would show ideally like where signs of potential illness popping up through crowdsourcing, but much contact tracing that requires people to participate and right now you still have questions. People are still refusing to wear masks so I mean. The. The degree to which you get people to all agree to wear wearable. Seems extremely optimistic and then when you deal with things like public. It adds all sorts of other complications so an office. Could all agree to do it? But how do you? What do you do in the larger world? I think those are questions. It still hadn't been worked out because the systems are only as good as everyone else's reporting, and just to be clear that this is not a cheap option. This is not cheap solution. Right? Like how much does this thing go for? And how easy is it the by one? The pretty easy to buy, and they're not that cheap there four hundred dollars so. Yeah the falls line with with. Your standard good smartwatch or Apple Watch or thing like that, but you know it's made like titanium plastic on the interior and It's it feels nice, but that's a lot of money, and I think some people will really like it because it's convenient. Enter praying and not a watch Some of the researchers also pointed out that a doctor's and frontline workers don't like wearing rings because they're not good that the germs could get in there. LEXIE! Shoes she's wearing. It didn't like it because of swelling and for exercise, the ring didn't feel ideal was uncomfortable for her and you have to get fitted. You have to get a particular size on the ring. They send you a sizing kit, so if your size changes, that's not great. A watch is adjustable, so there's a lot of things that are weird about it, but I think. It opens interesting questions about what temperature could possibly do. On, wearables in I'm really curious, what will pops up after this? It seemed like from your experience that made you call it somewhat invisible and the data I mean how ultimately how useful is this data? Have you used it to change your life for because you sort of hinted that that you looked at the data? then. It hasn't really changed anything, but like ultimately is this useful? In terms of changing my life. No, because you're right. You know this reminds me of like the talk. I had with with Kevin Lynch on Apple. Watch and apple could be making a lot of decisions on this to why they're only doing. Certain elements of sleep tracking. You know they're just doing the bedtime. Wake up is they claim that the rest is not actionable? Now they call so it could be that apple isn't fully develop the rest of the tools to their liking. But I think that's true in terms of when you get sleep scores like on this. What do I really do with like that? I didn't rest well enough you. There's really not much you can. Do you try to get to bed earlier? Maybe try to take it easy. Me could try to like take on yourself, and that's what the APP recommends like. You know you're reading the scores. Great. Go do it today, or it's not great. Take care a little bit today and I think that's interesting. So in that sense did change the way I would perceive some days I go. Hey, I'm not a great readiness score. just be a little easy on myself. What I know that just waking up and just feeling like crap. Probably you know I think some of these things correlate with how you would normally feel. Feel anyhow, if you're a self aware, but I think the getting back to work thing, the bigger question which is like you know I hate to even leave with that in the story, but it's what people think about about the possible Kobe awareness. I can't yeah for me I'm I you know I? Pi- blood pressure. I'm not going to put myself at risk going out. Out in the world, even if there's a sliver of it and then if I don't know that the APP, the rest of the world is behaving in a responsible way. Then I don't want to put myself there and that doesn't have to do other a wearing ring. It's like so the hard part. There was a halo over everyone's head. That said you know yes I. I am using the device I. Am I am part of your network? Then that's bubbling I. think that would be. That might be a different story, but again. None of these data things that these research programmes are absolute yet. These are all experiments, and all the researchers things a tip of the iceberg, so look how long people have been researching sleep and possible signs of. OF APNEA. or All. These research programmes with wearables they can go on for years and the NBA is very much an experiment. We don't know at all how that's going to turn out
Leaf Botany - Shape
"It comes to leave talk and God knows I love Lee Talk. There are a little terms. Bandied around that you may or may not be familiar with. I'm going to run through some of them now just to give you an idea of the range of terminology that you can use to describe leaves. I mean why bother you can say well. Relief is round or it's pointy or it's holy. Why bother with all these specialists terms well part of the reason? Is that as you get more into this hobby? No doubt you'll start reading up online and in books about plants and you'll find these terms start to come up and learning. These terms just helps to enhance your understanding of what you're reading. So what are the some of the terms that you're likely to come across? Well let's start at the very basic level with the leaf walk makes up the structure of the leaf. Well the Lamma is the blade of the leaf. The flatbed the that we possibly most interested in and the stock he bit well. That's the patio but do remember not who plants have patios. Some of them joined straight onto the stem. And it's an adaptation that saves the plant some water and listen to bobby reminded me of another useful pair of words when describing leaves and that's back. Co and Adak seal yet. You have to have your teeth in when you say those the opposite of visit belief. That's the Adak seal and the underside. That's the AB axial again. You might come across that one when you are reading about plants and that just helps you to know what is what. And then there's a whole set of words just describing the shape of a leaf. I think you notice about relief. When you're looking at is is it. Simple or compound now simple while that's fairly obvious it just means the leaf is one whole thing together rather than having some complex design whereas a compound leaf well that's formed by a number of flits that join together and then attach onto the stem and there's a couple of different coins compound leads your probably going to come across in the House plan world. Probably the most notable is compound pommie now and as is often the case the clues in the name a compound Paul Mate. Leave looks a bit like a hand. So if you think of a horse chestnut leave or in the House Plant World Shuffler relief. You are along the right lines and you can. Of course get palm. Eight simple leaves think spicier Japonica for example. That's a great example of a leaf. It looks like a hand but it's simple. It's all one leaf. The other form of compound leaf ease the P. natively compound leave and I guess the best example I can think of this. One is the sensitive plant Mimosa. Puteh Co where the leaflets are. All arranged in a straight line out in the garden the best example probably is arose some of the names of quite poetic. Iran the like Peltier eight which means a leaf where the patio joins not the edge of the leaf but somewhere in the center like Mr Shanley compete to that I also like has state which means a spear shaped leaf so think of a Philodendron artem being the perfect example again the Latin telling you something about relief and then we have the wonderful Lancia late which means quite simply shaped like allowance so in other words it comes to a point at the end. So think about your busy Lizzie. Impatiens classic Lawns Hiller leaves there are loads. More LINEAR world best. The spider plants leads right. Best fairly obvious. And then you've got something like Hoya carrier with its OPD coordinator leaves which means that harsh eight with the stem at the pointy and rather than the other end. If you want to go deeper into leaf shape names then do check out the show notes. Broil include some links to some wonderful pictures and diagrams of different types of shapes and he can spend hours learning the all. But how is a leafs shape determined? Why is this so much variation? Well this was where I need to call in an expert. I'm Enrico Coen. I'm research scientist Jonas Center Well we try and study and understand how plant forms are produced. How leaves grow how flowers get shapes often look at around my growing collection of House Plaza? Just wonder the amazing variety of leave shapes demonstrated even in my body's collection do have any insolent for us about why certain leaves are shaped as they are what what is it. What are the factors that determined the shaper indeed leaf size leaves a fascinating terms of the as you say variety of shapes the produced and one of the big questions which we still don't know all the answers to is how these shapes generated manmade shapes? We have a notion of how we make a spoon or plate because there's next to hand off our own hand the the process but with a leaf as with most biological structures there is no external hand it all has to figure out how to produce these shapes internally and Just as in a sense you could imagine how just to we use the musical scale. A single musical scale can produce all the different music that we hear from symphony used concertos to pop music and yet it's the same notes. It's the way in which the organized and put together the generates this amazing Variety of music and the same way leaves have a set of basic ingredients. And it's the combination of these ingredients allows still many different forms to be generated so although he's Mabul variety forms underlying that Some basic rules to get combined and also some glorious ways to produce the shapes. We
"research scientist" Discussed on Papa Phd Podcast
"You lived through that what feelings that elicited and how happy you were with your new new setting. Yeah I mean I think for a One of the things that became apparent very early on and I think is still a case. Is the high pace nature of being an industry? Setting You know I had great on boarding into companies are very high paced Environment and I really enjoyed that So it was a case of like especially with onto a chemistry background is like I get stuck in the lab day if he thinks For me it was very much like you go in and decent some some of this research to see what you find and I've been very lucky that I'm I'm I'm allowed to have a free rein over what they do and that was apparent. Very early on So for me like it was quite freeing in a sense that like I get to drive in the direction I want to An that became s dot. Freedom became very obvious. Very early on for me One thing I didn't ask an I should've maybe asked at the outset. Is what position were you hired into. What what is it that you're that you're doing now? What what what's the name of the position in was? What is it that you do in your day today? So I'm a research scientist. So I primarily work on the research and of research development and I developed new on. It's cool tools for the water industry and that involves see working with researchers from all over the globe we have we have open innovation Where we can speak to new people were you looking for the next the next cool thing so to speak. So that's kind of the role and when we were off the Mike you mentioned teamwork you said you you were talking about this culture of Teamwork in industry and I had to me when we were talking that it's something that surprise you something that you appreciate a lot. Can you talk a little bit about that? Yeah I ain't for me One of the things about about academia is that I one of the things that I found difficult is the competitive nature of academia gem ruin we know from from some research has been done the a large portion of of never go on to be a professor in this that kind of kind of competition often results in people not necessarily being collaborative and making sure that they're going to be the professor one day you and I. I really work that way. Like I like to collaborate and when to industry I found the the coversation was was great and the way more about the teamwork than is Abou individual success and I may be naive. Genera really buy into that like I really enjoy that kind of drive in the fact that we will put together we. We might have tight industry deadline. And we're like Kay what we're GONNA do and how we can get there. And how can we break this down so we get to that point and again just out of curiosity in in your lab and your run around you. Are there the people with journeys that looked like yours?.
"research scientist" Discussed on Papa Phd Podcast
"Part Two of my conversation with Zoe heirs. We discussed her experience. Transitioning from an industry led post talk to her current position in the water industry. We talked about the application and interviewing process in about best practices at this important stage about what you bring to the table as a candidate when you have ADT and also about those experienced so far as a research scientist when I go to industry I found that the collaboration was was great and the way more about the teamwork than it is about individual success and I maybe naively. I really buy into that like I really enjoy that kind of driving the fact that we will pull together we we might have a tight industrial deadline. And we're like okay. What are we going to do? And how are we going to get there? And how can we.
Humans will determine fate of Greenland's ice sheet
"Greenland an island more than three times. The size of Texas is largely covered by a massive sheet of ice. It's more than a mile thick in most places but as the climate warms the Greenland. Ice Sheet is starting to melt faster than it can be replenished by wintertime. Snow Twyla Moon is a research scientist at the Colorado based national snow and Ice Data Center. Where seeing more and more consistently years of very high ice loss something really that humanity has ever seen before and the trend could continue in a recent study researchers at the University of Alaska Fairbanks studied. What will happen? If carbon pollution continues to increase over this century. They found that within a thousand years the entire is she would likely melt causing cease to rise between seventeen and twenty three feet but moon says the world can avoid that future the difference between following a very aggressive path of action and reducing greenhouse gas emissions bursts. What we're doing today is quite truly the difference between losing all of the Greenland ice sheet or keeping the vast majority of it. That's all just a matter of Human Action. What we do will be the primary determinant of what things look like in the future
Bezos Hopes To Start Amazon Workers Coronavirus Testing 'Soon'
"Amazon CEO Jeff bays says vastly more testing for crown of iris is needed for the U. S. economy to re open and his company is equipping its own lab to potentially begin its own testing of all workers we have more on this from NPR's Alina Selyukh Amazon had reassigned some of its research scientists software engineers and other employees to team that's building quote incremental testing capacity for covert nineteen in a letter to shareholders CEO Jeff Bezos says the company hopes to begin testing small numbers of frontline employees soon bases in Amazon have said they're potentially looking to start regular checks for all employees but not sure how far this kind of testing will get in the relevant
What's the Difference Between Turtles and Tortoises?
"COM welcome to brainstorm a production of iheartradio. Hey brain stuff lauren. Boko bomb here at some point. Let's just say around two hundred and sixty million years ago earth got turtles. They looked strange. And these our modern mammalian times when lots of things are squishy and unarmored but during the late Permian epoch. Those early turtles were dressed in all the latest fashions. A short sturdy legs bony plates and a stiff splayed crawling strapped. Shortly after turtles made their evolutionary arrival eight fairly standard earth thing happened a mass extinction event. Although mass extinctions have happened with some regularity on our planet this one was a doozy and it wiped out almost all of the life in the oceans and over two thirds of the vertebrates on land. The things that survived had to have been pretty good at survival and it turns out turtles. Were we spoke about email with Laura Smith? A research scientist who specializes in herpetology at the Jones Center each way which is an organization in Newton Georgia that promotes excellence in natural resource management and conservation. She said turtles have a really successful body form. That hasn't changed all that much over time. They retained the primitive shell. Which is a really protective safe body design. Also turtles live in a lot of different habitats. Their aquatic and also terrestrial so living in a lot of different habitats has allowed them to persist. So what's the difference between tortoises and turtles all of the animals alive today that protect themselves with a Shell? Which is basically just a modified ribcage are in the order student. He's collectively we call this group of animals turtles but individually. We might call them different things based on where they live and some morphological and physiological traits. Tortoises are a group that are generally always found on land. Smith said they say that. Not all turtles tortoises but all tortoises are turtles be turtles are organisms with. Shell which might be in water or might be on land. A tortoise is a type of turtle in general both turtles and tortoises as well as other reptiles lay their eggs on land. It's what makes them different from Amphibians which need water for egg-laying end at least part of their life cycle because tortoises are a type of turtle. It's difficult to lay down hard and fast rules about what makes something tortoise ish rather than turtles but in general tortoises are always found on land whereas turtles can be found in aquatic or marine habitats as well as Land Smith said turtles and tortoises look different because of where they live a seat hurdle is only found in the ocean the females are the only ones that come on land and that's just eggs they have four legs but the front legs are almost like wings or paddles. They're not great for moving around on land at all because they're adapted for swimming quickly. Their shells have a low flat profile for cutting through the water. Compare that to a Galapagos Tortoise for example whose body can weigh up to nine hundred twenty pounds. That's almost four hundred twenty kilos with stocky elephantine legs a high dome. Shell and big scales on their exposed skin to protect them from predators and they wouldn't last long in the ocean but luckily they don't have to. The Smith said for the most part. There's not really one characteristic that tells you whether something is a tortoise or turtle but it's pretty clear if you see a little turtle on the side of the road and it has a sort of flattened shell profile webbed feet in the back smooth skin and some brighter colors. That's going to be a turtle tortoise. Have a heavier more dome? Shell and subdued colors as usual. The terminology can be confusing Box Turtles for instance which are widespread in the United States in Central America? And don't really swim or spend much time in the water. But they're still considered turtles rather than tortoises and then there are the terrapins. Which is the name given to aquatic turtles in the United Kingdom in the US aquatic turtles are just called hurdles with the exception of the diamondback terrapin which lives in brackish water in tidal marshes in the eastern United States? Both tortoises end hurdles have made themselves at home on this planet. We find both on every continent other than an Arctic with one exception. There are no tortoise species native to Australia. Smith said
Geometry-Aware Neural Rendering with Josh Tobin
"Art Everyone I am here in Vancouver at nerves. Twenty nine thousand nine hundred and I've got the pleasure being seated with Josh Tobin. Josh is a former research scientist at open. Ai and a co organizer of full deep learning Josh. Welcome to the trauma add podcasts. Thanks happy to be here awesome. So let's maybe jump into a little bit of your background. You spend some time at opening I and you did your PhD at Berkeley advised by Peter. Bill who's been on the podcast before? Yeah so I got into the field about four years ago then really. I started at Berkeley in the applied program. And then You know kind of on women took a class with Professor reveal who ended up becoming my adviser and I was just kind of blown away by the potential of Ai to make robotics work better and so through that. I ended up working with Peter and spending a bunch of time at opening. I and so. What are you currently up to yes. I left opening a few months ago and in the process of exploring kind of what's dive into next figuring out what's next yeah exactly. Nice Nice And so you're talk here actually later. This afternoon is on geometry aware neuro rendering. What's right what's that? Is that something you worked on? Ed Open or kind of personal interests. Yeah that's something I worked on at Okinawa. I also through like during my PhD at Berkeley Kind of While I was spending time in both places okay And so the kind of the core problem that I'm working on there is You know how do we look in robotics? In order to to act. Robots need to. I understand what's happening in the world and so typically the way that you'll do that in robotics is you'll take your sensor observations about the world and you'll map them to some way of representing the state of the world so where all the objects what poses are they in what configuration is the robot in etc but the challenge. Is that as you move to? More and more complex scenes than it becomes hard to write down a concrete representation of the state of the world. Nam The kind of the focus of this work is on implicit understanding. So can we. Can we take some sensor observations about a scene and then use that to create a representation of the world that Kind of implicit Lee has an understanding. What's happening the scene? And so what's Can you be more concrete than that or is there an example that That you describe that yeah absolutely so that concrete in the way that you formulate. The problem is You take one or more. Observations of the scene so think about cameras imaging the same scene from different viewpoints. And then you train a model and the goal. That model is to give him some arbitrary other viewpoint on that could be any other viewpoint of the scene. The models goal is to render what it thinks. The scene looks like from that viewpoint And so the industry if they can do that than it has an accurate kind of implied or implicit representation of what's happening in the the tuition rates if If the model can do that task well. It has to have some sort of representation internally. That understands what's going on in the scene practically when you do this academy cameras. Are we talking about yeah? It's A. It's a handful. I mean it's The the formulation can really extent any number In practice I usually used three cameras to image the scene. And then the fourth point is the one that the the model is trying to predict. That is the the neuro rendering part of the the talk. The geometry aware implies that it's kind of a model based approach in some sense. Yeah I mean so. The geometry aware part is the kind of the inspiration for the work that I did And what I built on this paper from deep mind call Geeky one Genitive query networks. Okay and so the the extension that I added to. That is the geometry aware part refers to is if you if you know The geometry of the scene so where the cameras are relative to one another than that allows you to constrain the process of of searching what pixels are most relevant to rendering a particular Pixel So you know if you're rendering a particular viewpoint and you WanNa know you know. Okay what should I imaging? This pixel somewhere in the the image that I'm wondering then you want to do. Is You want to go back. And Look at all the images that you've been giving US context and sort of search over those images for relevant information and if it turns out to be used geometry of the scene this thing from classical computer vision called the people are geometry then you can constrain that search to a line in each of the context viewpoints and so maybe as context kind of walk us through G. Qn and what that paper is talking about. Yeah so they're taking this problem of neural rendering rate so looking at looking at a scene for multiple points in rendering from an arbitrary other viewpoint and then they set up a model structure in the with that model structure works is basically an encoder decoder architecture. So the encoder takes each of the viewpoints and maps them through accomplish on their own network independently and so then you representation for each of those points. Those representations are summed and you get sort one representation for the entire scene. So that's kind of the encoder part like element wiser caffeinated. Yes some element wise. Okay Yeah So. Then you take that that representation for the scene. And he passed to the decoder And then the decoder shop is to take that representation and sort of go through this multi step process. Turn it into the what. He thinks the image from that viewpoint should look like okay and is there any particular intuition for the element wise some versus keeping everything around? You know I think the main intuition is that you want it to be You don't want to depend on the order that you present the Cameron So you can candidate than order matters. Some it doesn't okay. And so they they're results. Were purely based on this Just the camera angles in the kind of encoder decoder architecture. And so that when you added was a constraint to the search base that is based on what we know about the geometry. Epa people urgency polar geometry. So so so. Basically the the contribution of paper is You know there's this sort of bottleneck in the formulation where you know you've taken the encoder and that's giving you Representation of the scene and then pass that to a decoder. But what we do instead is Instead of having this representation of the scene instead we haven't attention mechanism. Is that attention? Mechanism allows you to when you're generating a new image to attend over all the permission in the in the representations at the context images that you've been given in the way that attention mechanism works is by taking advantage of this this sort of fact of three geometry called the the The geometry is calling it and attention. Mechanism is it. Is it like attention like or is it implemented the way attention is often implemented in these kinds of networks? Yeah so it's it's implemented exactly like what kind of like a scale products attention but the way that you have to do that is you know you have for every pixel. You're searching over a line. You need to create unique to construct the right br representation so that you can do attention in the normal way.
Racism at the school gate and education reclaimed (Part 2)
"Aikido got You Butterfly Net Enhancement. Gravity Heaven and what she step. We're right next to the Bank of Torrens River on Ghani country in the heart of Adelaide plays. Don't fall in the torrents. Those of you that are not South. Straighten off a good ribbon polling today on the white out feeling very sorry. Try and start wipers water Do into the podcast of last week. Show if you've arrived at the dorms late because we're on camp here on science friction and it's going well. Everyone's getting on. This is the C. Syros Aboriginal Summa School for Excellence in technology and science and I'm living in with the nearly forty indigenous students from across the strategy here from Perth to the Torres Strait. They've come from far and wide and it's great to have you with us to on the Tesha Mitchell. You've caught an enormous something or dragonfly. It's huge thing. It was just innocent. Thought Maybe we could Kate Me Swan and then if we get any more or less just Alana line Patched candidate the University of Adelaide so working on parasitic wasps. My mission is to teach a bit more of an appreciation awareness of what's around which these kids have embraced wholeheartedly fishnet Connor Looking. It's camouflaged really. Yeah so as part of ABC's walking together. I'm bringing you powerful personal stories from three generations of indigenous. Australians today on racism in classrooms on triumphantly pushing past the low expectations others can have foyer and are knowing who you Would Hi this is a Science Camp Theresa? Let's get some of that good stuff at by the reba without insect. Nate's I love it because when I was little I used to do this in the backyard. I'll just for the fun of it. Like we did. Ones and lacked playful the bugs and stuff and Done things we went touches. The real big because that's scary. This is year eleven student Catherine. She's from Queensland. I've always had a interest in biodiversity because when I was lying about it in school I just found it fascinating the way things like adapted to the surroundings and how strong Some animals off. But do you think you might study in Uni? I definitely want medicine like the medicine. Science and even in science medicine side of it because on surf fascinated about the way humans like animals too but mostly humans alphabrain the actual workings about nerves and our nervous system and everything. I just find it so fascinating to fix people with your knowledge of that. It's just it's mind blowing to me. If you're ever original or Torres Strait islander you make up about three percent over strides population but just under two percent of all students enrolled at university are indigenous. That's growing by around half of a said over the last decade or side when it comes to Unical says in the natural and physical sciences. It and engineering. Less than one percent of students are indigenous for first year medicine. That's around two point. Four percent and of course completion rights alarm but this camp is about helping to change that. It's about road tasting university. Simon names macaroni. I'm an epidemiologist with. Csiro food and nutrition and things are about to get very real for the students right now. We're talking about their activities for the rest of the week and in particular their inquiry which is quite a lot of pressure for them. They'll need to spend a lot of their time thinking about the question that they want to investigate for the next few days and then they'll have to be ready to present it by next week. You asking them to do scientific experiment in two days scientific inquiry. That might be an experiment but it might be some other activities but yet in today's Yep they'll spend a lot of the allison a day. Doing it will be under a lot of pressure but based on previous years they do a great job so they've got to collect data definitely have to collect data they'd go to interpret data and they'll go to present it all of the precious situations for them so the pressure is on from pretty much all mice now not quite a couple of days. I think they'll feel it from tomorrow morning. Hitler research can be conceded. A A dirty word Saith West head is a young research scientist irregular mentor on these caves. He comes from Alaba Coal. And we're edgy. Country in these half miles research was something that was done on. Aboriginal people not with Aboriginal people and certainly not let by aboriginal people but as we get more aboriginal academics in high positions within the academy. This is where we can start to see a change of the culture so we need young people. All of the students present curious and inquisitive mind and from my perspective. That's all you need to be a scientist. The rest is just learning the specific language to answer the specific questions that you come up with and that's just a process. Anybody can do that. We really made more indigenous people in science. We've got so much work today. But we need more indigenous people everywhere. It's hard to access education for aboriginal people and are stolen papal. It's hard to walk to welds of wanting to preserve your own culture and sense of identity. Sometimes studying integrate main sacrificing culture identity and sometimes staying strong culture means sacrificing education. Perhaps no one knows these more than an ano education later. I made it a gathering by the five page of the Wheelchair Boarding House. Where all staying at Miami's Ruben and direct for education does P. Y. Yeah we didn't on almost for you know all the people it's our language and then another language as we had last week students from the remote traditional lands of the unindo people in South Australia. Come stay here. We'll check to go to high school in Adelaide. Now looking at you know dairies. To Wolves do peak will come together. You know the wisden world is really important that are now people need to get educated through sure school to get a job and money travel around you know speak language English and understand where there was an will come from and why why we see really important pulled in you know kids to university by his crowd. I need to build than me on the stand with coming from and you're not educated to vision. It's a big thing you know for
Scalable and Maintainable Workflows at Lyft with Flyte
"My name is skate then I lead the flight team And probably one of the founding the founding person of light at lift so my background is I looked across different industries from hedge funds do retail logistics to cloud. I'm hoping And now finally back in sort of transportation area and one of the things I've been interested in is just large scale. Aw data processing solving business problems aware data at computation comes together and machine. Learning is a place where you know. That's really I started lifting twenty sixteen but I started flayed the close to the end of sixteen It was mostly like you know I I started working on this team We were trying to get ETA's which are At explain what Anita so when you open up lift APP and and you see hey three minutes to get your driver are if you are in your car and you see like it'll take fifty minutes to reach the airport. Sometimes it accurate. It's mostly because of a ton of machine learning models that go in the background and including understanding how the roads traffic is and understanding. All kinds of I think that are happening on in the current conditions on the road so using so I was leading that team and Joined the team. There was another engineer on it. He used to run the models all on his laptop and he he running this and he's like I had the script. I've just run this and it just ended figures out and then this other script tripped this other script and dude. That's crazy so I'm trying to decide whether interrupt you and just like dive deep into that. That sounds crazy. It is a model. That's doing like live prediction of ETA's training training the law model uh-huh collecting the data for the model. And things like that right and I still reproducibility as us and stuff like that. A lot of issues. You wouldn't be surprised how many times this happens in the industry just like the and and this is actually how we lead into it because this is the current state of The infrastructure for machine learning earning especially production models didn't exist. We didn't think about retraining. These models at that time and quickly we wanted to retrain them and then the laptops opts not gonNA kill the other story. That happened So this is just leading into flight right and the other stories that happened is devoted research. Scientists have my team He created a model models pretty cool. It's or lift for many years But he left the company in the model went with him. Probably heavy last epic we had no idea of what and As leader told me. Hey let's recreate this model. And I don't know how to and we knew the algorithm so we just rotate. You know got everything done. It would not give the same results And it like it literally took US US three months to get like the same level of accuracy. And we're like okay so he had done all of this extra work really kind of lost. We didn't waste wastes too much effort on it because we knew we knew some of the tricks but still it's wicked effort in trying something out in going and trying out the accuracy and you're like Muslims mental. Yeah Yeah it's it's not like one person spending all the. It's it's wasted effort so at that time we decided to do S- needs to this needs to change and delivering new models became slower and slower. So that was the birth of flight. At that point. We used to call it a bad name. I'm going to put it on the podcast but he used to like I wrote a first draft proposal internally. The everybody was like you're crazy. This thing is not gonNA work But somehow in like a couple months of you erected a we we one of this thing and we got a team to try it out and and This team was also struggling a lot with delivering their models. The intersection where flight really fits in is when you have a lot of data anyone who produce repr- reproduced your models again and again like maybe every day every week or every hour and you want like the trace of what happened in the lineage between everything and this team actually fit the bill And they for the first first time they were able to deliver a model in like six months and this was a gigantic marlet affected the bottom line of lived and it was really really meaningful and it was not without a lot of breath. You know stress than working hard at through the night but that was the starting of flight and that was in twenty seventeen and and then region stop like the the use of the company just skyrocketed And and at that point you're like hey we should open source this thing because because it's such a big problem to solve that small team it live can probably never solve it on their own. If you open source of me should be able to work with the community be here more ideas and improve it all the time so we actually wrote the Everything from scratch made it Kuban. At his native took Like the primitives that we understood from looking at all levels use cases. I'm decimating amazing. Part of lift like it's a rich ground of amazing use cases and we used all of that put like basically distill that information into flight. And that's our first cook Into the world them. Tell us about your Your background in what you do lift shore. Yes so my name is Heison to I I have worked previously at Microsoft Bingo and they had a journey. You know up and down the stack a working enterprise great occasions in Louisville's stories stories And they you know at some point. wanted to try out a meal in light kind of found this sweet spot spots in the email infra to fit the bill. Kind of thing for me is doing two years ago January and and at that time it was that we're still stabilizing the international flight unnamed product It was great and teams loved it but as Yeah Stephen Seeing at that point I joined wind in the midst of this Discussion about what do we do next So I got to be part of the decision Zimbabwean going Cubans native and All the the icy as critical design design decisions we We took in flight light to officially decisions. All the the league going With a very strong type system very strong language specifications through Protopopov. Love like there is a lot of things. We view We are very opinionated about in-flight and a lot of things we are not we are. We'd like explicitly decided decided to leave then We based on experience. We had we think we found a good pass for where we like the give you the learnings like enforce the learnings we have had before In how we ask you to write your goods or deliver your Morton's or the processing tasks or whatever and the same time. Leave it open for a variety of different workloads. That can run the system and here. We are in very proud with how the product turned out to be in the lawns and the reception we have had during the conference prince and Just a shout out to the team wouldn't have been possible without like crazy amount of effort with the team. It's an amazing team. It live And we are proud if all our users also at lift just stayed with us through. I'M GONNA go times and and thank you for all the support. It's awesome awesome so okay then you've given us a kind of a little bit of an overview of light. Maybe you take a step back and you know what's the core value proposition that A flight is offering and hide them. You mentioned that it's communities native. Like how does it relate to to Q.. Flow for example. The fission Let let me start with the motivation. Like or what is it that we think is missing and what we were trying to address one of the things as I said we started in thirty seventeen. So that's like the landscape was very different at that point rate. So we've heard from that point and this is a we to even though actually I think this is the real one but this is a we do so that means we went through a process of like actually making something and failing and then redoing it. That has a lot of learnings with it. So so one of the learnings that we feel that there is this artificial divide. That's happening between data. But actually they. They go hand in hand tonight lake you. It's not that these companies have amazing data system. They're not the googles. facebooks are the Amazons of the world right they are smaller companies nimble. They want They are basically pulling the data stack to So and the other thing that we realize is their teams cross collaborate quite a bit machine learning models ability by a team but the team be probably provides the data that bills that machine learning model and actually the the fallacy of Separating them is that many times in production we use machine learning models to predict ached and that creates data that becomes a fact in the fact tables in the data would and they made aims use machine learning models to connote that factor dimension and which drains other models so this cyclic nature. That's happening and this needs to be captured at that. granularity saying that you know there is data and processing and and machine learning all interacting together so that was the motivation behind flight that we need a single tool and a platform that allows for collaborating collaborating sharing and Emma lops along with And with with definite focus on arcusfoundation. And that's why I had. The core of flight is is a workflow. Engine that actually runs all of these pipelines but from the point of view it was built for collaboration and sharing across the company various aspects specs as well as the processing and Machine learning on the same tool
MIT CISR Principal Research Scientist
"Which genie Ross welcomed. It's great to speak with you today. Thanks for having me. Peter Ought to pleasure. I've been looking forward to this conversation Jeannie many people who are listening to this will know you as a The former director of the MIT he sloan's schools center for Information Systems. Research or SCISSOR as it's referred to as an organization that you've been with her for twenty seven years now you're the principal research. Search scientists have the organization also multiple time author of it. GOVERNANCE OF ENTERPRISE ARCHITECTURE IS STRATEGY TO BOOKS THAT I've recommended many many times. And and now a new book designed for digital how to architect Your Business for sustained success and I thought we begin there The topic of digital is pervasive basis. Certainly on the lips of of of most executives I would say let's say many executives in every industry and A topic that is so. So broad The topic of digital. I wonder if we could begin with that definition How how do you think about the topic of digital how much of it is truly? We knew And how much of it is perhaps a new brand name of some topics that are older. That's a great question to start with. I actually will take you back to the start of our research. We have one hundred global sponsors of our research center and we just tried to stay in touch with with what's going on in their lives and about five years ago now we talked to mostly. CIO's these organizations and said what's going on in in your world and our summary of what they were going through as they were being bombarded by these digital technologies it was social mobile and another cloud Internet of things artificial intelligence blockchain. And you know these technologies just kept coming at them and with each one. There was some more hype about. If you don't start using this you're gonna miss something and our sense was that CIO's where we're being instructed to make good use of these technologies when their companies as we're still struggling to use information technology efficiently you know they're still trying to get their ERP's and their crm's to produce real value you and now we have this whole other layer of digital technologies. So we want to study and I think when people talk about digital well the thing to bear in mind is just what all these technologies are making possible. That wasn't possible before we have had information technology ready for a long long time but now we have technologies that are making data ubiquitous. They're making connectivity constant and and they're offering massive processing power and and the reason. I think there's so much excitement. Is that on the one hand they let us do everything. We've always done better her. We can become more operationally excellent but they're also allowing us to do things we could never have imagined. That was the Uber Moment. Where where you can do more than give somebody a ride? You can give them information. And I I think that is the digital phenomenon the recognition that things are possible that were possible before very interesting and the whole notion of designed for digital as I as I think of that title there are two different camps very different. There are those organizations born in recent years who were designed digital from the get-go that they are born in the digital age the digital native and as a result of that they don't have at least at the time of creation the burdens of legacy Then you've got the older vintage organizations and what is sometimes referred to as the digital immigrants Those who were born years ago multiple decades ago before the digital age that the digital age is you just defined it and they have legacy versions of people practices and culture a process certainly technology in an increasing debt bear that needs to be retired and in order for the latter camp to better compete with the former camp. There's a significant transformation. That's necessary And I know of course that a lot of your research the companies that you're that you've been nearly two hundred companies in fact that you collaborated with in order to draw the insights for this book are more from the ladder campaign from the former. But I wonder wonder if you could talk a bit about that that that challenged that these organizations have in competing with with with companies that can move so rapidly as a result of not having the kind end of anchor weighing the back. That is that legacy. Yeah you know I. I think it's been overstated that we're going to compete with these startups APPs. We've certainly seen it. It has happened But my sense is that The big old companies. This is my affectionate terminology for successful companies. Big Old They have some real advantages as well as a legacy that sometimes gets in their way they actually know how to scale for example they they know how to complete all their transactions very efficiently. A lot of the startups will end up struggling with the basics so their their moment and to be able to move fast and not having to worry about legacy may in fact go away pretty quickly so I think we probably overstate the disadvantage disadvantage of being big. An old in a world that You have to move very fast and you have to understand the opportunities but that said I do think I think that a big old company that wants to be successful in the digital economy does have to learn how to do things quickly. How to Experiment with new opportunities to learn what's going to work and what's not going to work with customers and I think this is something that the digital startups have tended to do naturally They every startup. We studied Has this history of saying well when we started. This was our idea where we are now is somewhere pretty different. I often point to AIRBNB for that. You know the initial idea is will throw a mattress and air matches on. Somebody's living room floor lar. And they will Service breakfast in the morning and then we're going to become friends and they're gonNA show us our city and it's going to be a real community building kind of environment and and you know that just wasn't of interest to that many people so they saw what was possible and they adapted and their strategy kind of seized. Just the opportunity that was there and the thing. I think that will be hard. And it's proving to be hard for established companies. Is this idea that you can't just pick a strategy and insists it will succeed. You have to try a strategy and see if it will succeed and if it doesn't pivot and I think that's a real challenge for big old companies. But if they start to learn to do that they get very excited. And then they have all the advantages advantages of brand name recognition and and the ability to scale that gives them a real edge on the startup.
Social Intelligence with Blaise Aguera y Arcas
"All right everyone still here in in Vancouver at Noor ups continuing our coverage of this incredible conference and I've got the pleasure of being seated with bless Aguado. yuccas blesses is a distinguished scientist with Google. Ai Bless welcome to the Tomo podcast. Thank you so much. Thanks for having me absolutely so you are doing an invited. Talk here at the conference tomorrow morning on Social Intelligence and we're going to dig into what exactly that means for you but before we do love to get a bit of your background sure sure so It's a little motley. I started off in physics undergraduate at Princeton and I studied physics and applied math. There I I took a year off between my third and fourth years because I was not a very good student and I really started to get into into biophysics this X.. Pretty heavily so you're euro for after during during a or a little bit a little bit before and then during I worked for for a little while while in there he was working on bacterial Metaxas. That actually gonNA figure a little bit into mytalk tomorrow morning. So it's the behaviors years of of the intelligent behaviors of bacteria. And how does that. They that they find food. There obviously a really small simple system but maybe not quite as simple as people think okay and end and then from there My my next adviser Bill Bialik is somebody with a physics background. As well but also computational neuroscientist. He ran this course in woods. Hole at the marine. Biological lab called methods and computational neuroscientists it methods and computational science I don't I don't know if you're familiar or how many of your listeners are with with with them deal with marine biological laboratory but it's this place where a lot of Princeton notes on Cape Cod. Okay and so. It's right on the elbow of Kit. Kat across from Martha's vineyard okay this this little tiny town. It's very cute. And there's this kind of ramshackle lab that's been there since the nineteenth century tree that That a lot of a lot of visiting Sort of neuroscientists and biologists have been going for many many years A lot of really basic basic discoveries in science where made their. Oh so it's kind of this cool place. And and at this. Course at nothing computational neuroscience I I met my now wife Adrian Hill. Oh so she also came up in physics and Studied originally chaos and turbulence and fluid dynamics comics and things like this and was making the switch to puck additional science so we met there and and then she ended up getting a faculty job at University of Washington which is how we ended up moving to Seattle and around that time I started a company And was no longer really sleep. Part of academia at that point and the company got acquired by Microsoft couple of years later and they come into doing computer vision type of work or a it's a somewhat somewhat doing sort of multi resolution representations of of documents of of various kinds. It was okay. It was a combination of wave. Latouche kind of tricks six and and you X.. If I think wave letters like Kryptonite for me that was the hardest thing that I studied in Grad School. For whatever reason it was very difficult to rock it was it was hard. Yeah my my advisor. In Grad School in applied math was ingrid do bitchy who was one of the inventors intercept wavelength. Yeah she was she was absolutely wonderful very very smart very kind and I think I think one of the greatest living mathematicians if I. I don't know maybe unbiased. But Anyway Yeah Microsoft acquired it and I did immediately turn the team toward more four kind of computer vision e things right after that so photosynthesis which started off the photo tourism project by University of Washington professor and Microsoft research scientists together. With with their Grad student snively was in three D. reconstructions environments from the images and that was really my introduction to computer vision Asian. That was pretty classical. Wasn't like deep nuts or anything like this geometric computer vision but I kind of fell in love with that with that field and ended up at Microsoft Echo soft. You know sort of doing a lot of leading teams doing that kind of work so Microsoft's OCR team and they're kind of photographic treat type teams the teams that ended up doing a lot of work for a hollow Lens tracking The head using our facing cameras. All that kind of stuff was okay was part of my team at the time so I was at Microsoft for seven years I also was the CTO of bing maps which also had some kind of computer vision? The are photographic tree kind of stuff going on and being mobile and then I am I went to Google. That was six years ago. I come across so many people that are in this field that have some connection to bang. Yeah I shouldn't I shouldn't Bad I mean it was it was it was creative and scrappy at the time You know whether whether Microsoft was really committed to running these things I guess it. It's anybody's guess right but but yeah. I mean one of the most one of the reasons that I ended up leaving Microsoft was because about six years ago they had just Kind of lost the phone phone war and it became clear that they were going to be moving away from being a consumer focused company. We're GONNA start working on just enterprise stuff and I wasn't that interesting to me and that was around the same time. I'm also that did the whole deep learning revolution was really getting into full. Swing and I was very excited about about some of machine learning and computational neuroscience verging and Google is the obvious place. where the kind of hotbed of of a lot of that so nice? So what do you research. Google well at Google I started a team. UNCALLED CEREBRAL. With is not a name that we've generally used in public but that's not at all heading. Thank you it's the plural of brain. So there was a brain team already that you know Jeff. Jeff Teens started years before and I went to Google to start a team that would take how much more decentralized approach so rather than one brain. It'd be many brains. Everybody would have a little brain and I had a very augmentation focused point of view. You know the rather than having one giant running in a data center these things would have to shrink to democratize. There would have to go into devices. Run locally I had a lot of reasons for wanting to push in that direction including privacy Which I will talk about a bit tomorrow so mobile nets and a lot of these kind of efficient ways of running neural nets locally came from From our team again. I'm running. The you know the the the groups At Google the two things like oh CR and face recognition and a bunch of other sort of image understanding Primitives but we also power a lot of a lot of a or features chores or whatever you WANNA call them in android and also on other kinds of devices include including these little coral boards which are sort of an Iot kit for doing taking local I think those are just well. I guess it's maybe half a year ago at the developer conference drink. I have one. That's that's right that's right so yeah we're very excited about those cool he you mentioned OCR and Of all the things that we've talked about I think of that or it's probably easy easy to think of that as a solved problem the problem. But there's probably a lot of Even saying it. There's probably like this last mile problem. Where in order to get to usable or better levels of Accuracy and performance kind of that those last few percentage points are are really hard to get to. So you say I mean it solves problem and yeah I mean. It's good enough for practical use engines. That are good enough for practical use but a of of course. Extra percentage points are always useful. A little bit more is always better but also a team that I run at Microsoft was still using a lot of these classical techniques that would I you know they'll have a whole pipeline of different stages first segmenting out letters and then you know doing template matching and then using language modeling all kinds of like this and the direction that that that I think in the end that the you know the people in the team believe are really the most fruitful now are much more and much more neural so imagine smoke scanner that scans the entire line maybe by directionally and emits a string of characters. Kind of like a speech engine. Might if you you do it that way then you know. Join join letters and ligature is. Don't matter right cursive doesn't matter handwriting. And you don't print could be the same Arabic and other languages. That don't have good distinctions between letters. I ain't going but rather that don't that don't distinguish clearly between letters in the more cursive sort of approach. All of those things work and that sort of general and also just weird funds. There are a lot of things that are easy for us to read that a classic engine right so thinking about it more like a real vision problem some of the brain behind it as opposed to just a classical kind of letter clustering problem with the language model talked on
Music & AI with Pablo Samuel Castro of Google Brain
"Am here at nerves continuing my coverage and conversations from the thirty third nerves conference and I am seated with Pablo awesome. Well Castro who is a staff research software developer at Google Pablo. Welcome to the PODCAST. Thank you thank you very much for having me. This is a real pleasure to be here. Awesome thanks so much. I am really excited to jump into this conversation. You are someone that I follow on twitter. And like we've had these kind of back in occasional occasional back in overtime and it's great to finally meet you in person. you've got some pretty varied interests You spend a lot of time. You're research focus on reinforcement learning. You also tweet a lot about music and arts. Looking at your background you've done applied l. l. stuff at Google on ad from and other things you know. Tell us the story like how to all these threads come together So well originally. I'm from Ecuador and they moved to Canada after high school to to come study at McGill So eventually I did. My undergrad in the navy actually actually did my masters and PhD St at McGill with throwing a pre-cup and garden and so part of the reason why stayed in Montreal and McGill was for personal reasons. I had I was dating someone. WHO's now my wife and I also yes and I also had a band so I've always always been heavily involved with music? I grew up with music. Learning music. Play music so that was very important to me and I didn't WanNa leave that so I decided to make that choice. I know it's not typical thing that suggested to do while you're in the same university but for me. It was more important to to play music so I graduated. Did I finished my PhD at around two thousand eleven and then I moved to Paris post doc and this was at a time where a isn't what we see here with twelve thousand people in this conference. In Europe's didn't have back. Then it was called Nips maybe four thousand people So I WANNA say Nakajima and I was working at the intersection that was very theoretical between between Markov decision processes and form over vacation so I was finding it really hard to find a job because I wasn't former fixation enough for the former vacation community and I was I wasn't reinforcement learning enough for the reinforcement learning. Okay and so after my post doc I I just feared already have two young kids. And if you're that I would speak going post the post stop for too long so I luckily got up from from Google doing applied machine learning and adds an extra said goodbye. Getting at that point I stopped reading papers and faster fast. Then I did a little quick stint in chrome doing a building machine learning infrastructure so backend infrastructure And Brain opened up in Montreal and mark. Belmar are who I had done. My masters with he was he kept in research. He was in decline for a while and he was one of the first people to join brain in Montreal and he put in a good word for me and So then they. They offered me to join them and I jumped up that possibility and I hadn't been following the research. That also is a huge shock to come back. I I mean when I was doing my research. We were all working on Grit worlds and in pretty simple environment because a lot of it was theoretical. We didn't really use deep networks at all for enforcement or any now so it was a lot of catchup trying to to familiarize myself with the literature and how the whole landscape has changed so throughout all this time I always kept with music. I had a a few different bands. Always I've always been performing live and writing music and The other thing is when I started my PhD. I was actually considering doing a PhD with Douglas AC as as well as with During a pre cup in something with machine learning and music but at the time the what was available for music generation didn't really excite me very much Because it was still in the early days and I fear that it would taint my love of music and I just want to keep my music site separate but when a rejoined the research world and I saw with the Magenta team was doing I was kind of blown away by by the quality of of things then. I decided to also start going along that pathway pretty almost I think the day after I joined brain This artists from Canada. He's called David Usher. He's pretty well known in Canada. He approached us wanting to the other. He approached us that he was actually. I had abandoned the nineties called Moist and really popular and and he approached us. He wanted to do an album using like ai techniques and so we just Matton Kinda brainstorm then thing. He gravitated towards the most was lyrics and and So Google who was my manager at the time was Very generous because I had just joined bright. And he's like. Do you want to take this project because I like music as it sure. That sounds fun. I had never trained a language model. We're still trying to figure out all the steep networks because I hadn't looked at that but yeah google that gave me that opportunity and and I learned a ton and that project it's still it's still an ongoing project. So relative to the first model trained with David which we actually made a video out of that like he wrote one of his songs with the first prototype and it worked okay but the model we have now is so much better and I understand all of this language modeling so much better than they did before. And that's just ah that experience kind of showed me to not be afraid of stepping out of because even with reinforcement learning which is the background to step out of that comfort zone and go into two other areas that I'm not as familiar with because they're all interesting problems and really trying to dig into the details. And for me the way I learned the most is actually actually trying to implement some of these models architectures and play around with him because you read about them in papers and you kind of get it fine but until you're actually trying to get it to work for yourself it's that's a whole different experience and I've learned so much just from doing this like jumping from a one to the next in a separate can field and learning about those architectures architecture's but while still maintaining my research and reinforcement learning. Well it sounds like you've landed in an incredible place to do that. Not just kind of the resources of Google and the people that you're surrounded with and have an opportunity to interact with but your role seems to be defined as like advancing research. You know the implementation absolutely. Yeah so I'm a software developer like. That's my official title. There's also research scientists that Google and until recently there was still like most people that are in research wants to be research scientists. Because that's like then you're officially doing science So my like if I had graduated say four years after when I graduated likely would have been applying for research research scientist role Back when I google. That wasn't really a maybe Sammy. Benji was a research scientist but probably about it And so I entered Google ads syringe India and sort of advance my career in that in that track and when I joined Google it was a software engineer. Develop developing comebacks. 'CAUSE engineer you get an iron ringing. I don't have that Initially I was a little skeptical because the official description is your. They're more supporting research. Scientists and so. I was worried that I wouldn't don't have the flexibility to pursue my own research interests. But it's been not at all like that so I lead my own research projects and I still support a lot of people with the engineering aspects of it. Because I've been working on this a lot so I'm more familiar with Google infrastructure and just coating in general And it's been a lot of the major major advances that we see in machine learning the I nowadays a lot of his engineering. So there's of course there's still math and there's still a lot of theory behind it but a lot of engineering and and I don't think it I think more and more it is but Few years ago I don't feel like dot the credited. It really deserved and so living in the sort of intersection of of pure engineering and pure research is for me super exciting because I kinda get the playground in both worlds and learn from both when I've got a a long Melissa things that I wanNA talk to you about but you mentioned Something that's got me really curious. The you know what it means to evolve a language model so you started this project with David And came out with this early crappy language model and have evolved over some number of been like uh-huh Yeah No. It's been like a year and a half it's been or actually it's been almost like two years. I think since we started it but two years calendar calendar wise. But but it's not it's not one of my main project so yeah exactly so it's when I get a chance that I that I work. Yeah so as I said when I started this project I had never trained a language model. I like like I knew what else were studied in school. But so the first thing I did was I actually Andrea Sherr potty has the Yeah this famous blog post host The surprising reliability of of recur neural networks. Something like that thing. Anyway that blog posts and they got his Kodansha Jordan's are played around with it and that was the the Vero model. I'm just over characters and then I started tweeting that a bit and and finding new data sets for lyrics and that initial model that was basically a variant of Parties model was the initial model that I had and so that was okay. They just a milestone like okay was able to train. This actually get it to do what I wanted to do. But obviously was Has All the shortcomings that these types of models do the around around. I mean the the tension is all you need. Paper had come out not not Not Too much before then. And so then I started looking into these attention models and and so so it seemed like the right thing to switched over to to the transformer model and started playing around with that and so the V.. Two model was attention model and it's had various versions of a two part of the difficulty that had with the language with training. These language models on lyric status at is that the lyrics said is not the best in what sense so the tricky thing about these language models is that an end for lyrics in particular is that you're trying signed to get this model to learn English kind of so how how to structure English phrases together but in quote unquote poetic way and to not be boring doing right because you're trying to use it for creative purposes and you don't want it to be boring so we train this model and if you look at it like perplexity scores and things like that it was doing pretty well on this lyric status but but then when you actually look at the output. It was extremely boring so because in pop songs you have lines that repeat often. I mean that's just how songs written so the model would tend to just repeat the same thing over and over and over and It also had certain phrases that would keep on coming back to just had very high likelihood so I wonder if you've talked about this. I say like it's Hanway but the average pop line over the last six decades is you know that I'm the one and That one came up a lot and you can also get you know that I'm the one baby. So that's the average pop line. It was boring and so the interesting thing about working with with with David is that I build variants of these models and nitro him and one of the things he remarked on. Is that It was very nonspecific in the sense that at the nouns that it was using it wouldn't use proper nouns. So would you like me. You he she they since very kind of ambiguous. If you think of Like the Beatles mister mustard polythene pam jude. You know there's all these I mean the fictional characters but they're very canvas and so then you can sort of the ground the song song in something kind of real whereas if you're just talking about him like hey you don't don't even though pink. Floyd has a hate us
"research scientist" Discussed on IT Visionaries
"research scientist" Discussed on KOMO
"As a research scientist for Washington fish and wildlife he says this new amendment about killing sea lions doesn't apply to Puget Sound but that the state is trying to find solutions to fix this growing problem all of the you do sound is dependent on salmon to some degree best Cammarata Swensen Puget Sound and encounters seals frequently when an animal pops its head up and stares at you it's hard to not to associate that with just meaningful life she's conflicted about the possibility of killing seals and sea lions in Puget Sound to protect CNN it's hard to see them and engage with them and have it have a look at them and not sink that that's wrong Washington fish and wildlife is working with local tribes to identify any hot spots where senator killed more frequently find the potential of killing or hazy marine mammals to preserve local salmon this is an incredibly complex and controversial issue as well it's important to note that same an arch is struggling due to see lines and seals eighty man it's related to other factors as well for example warming waters as well as a loss of habitat Abby a county companies know is accepting public comments about the new rule on killing sea lions in the Columbia River comedies time is now for forty eight starting the new year in a new school can be overwhelming for kids but hundreds of students did together Wednesday morning in Silverdale commercially stole was there when the remodeled central Kitsap high school reopened. students can now look straight through the floor to ceiling windows taking advantage of one of the best views in Silverdale I think the schools great they've done a great job building for us I came in middle of August for more than my ASP meetings and we got to like behind the scenes tour and just the whole time my whole crew was excited and laughing and smiling grads like myself remember taking buses all the way across town to Olympic high school for tennis practice and home football games but now there's a whole stadium to play home games at home not to mention tennis courts and a baseball and softball diamond hill for the doors and come in and the second set of doors we don't we're not turned on here yet but that second set is locked and your only access to the school is through the office the new auditorium has a sunken orchestra pit and seats for nine hundred people a big difference from the school that's been here since nineteen forty two the hope is families filled with generations of C. K. grads who walked these halls now have a reason to come back would really like to have music of Vance summer camps you name it we want people out here in Silverdale I'm least all come on news United Airlines offers a solution for travelers who may be hesitant to step on board a seven thirty seven MAX jet once they're cleared to fly again the airline announced yesterday that fees we way for people wishing to change flights the plane's been grounded since March after being involved in two deadly crashes the be allowed to fly again once the FAA approves software upgrades..
"research scientist" Discussed on This Week in Machine Learning & AI
"There's something around container. So we were baking stuff, a lot into the EMI into the machines themselves, when they're started you're just directly because not everybody was familiar with Docker, but we picked up Docker too, because there's obvious reproducibility benefits. And when you have a lot of things quickly at the beginning of a research project having these kind of Docker file where people can reproduce you environments and not just your, your, your experiments, that's actually extremely helpful for collaboration in team. So we also ties back to that agility, and being able to move quickly booting up. Machines. Yes. And our IT folks, we're so happy because it's not like this doesn't work, but because you to get installed something that Rex this and of course, so up. So really embracing ups even for researchers actually was quite powerful, because you can only do the research, that's, you know, the mastery of the tools is really important to and part of the research beyond, you know, just Jupiter notebook. Let's say it's awesome tool. But if you want to go beyond, you need to master other tools. And that's that's what we've been doing. That's a journey through engineering craftsmanship as much as deep learning research, the, you talk about applying DevOps in this world to what degree in your experience. Apply directly or are there gaps or it only takes you so far, you have to modify the way you think about it. And I realized that I'm saying that as if DevOps is this well-defined thing. I think it's a good question. I think there's like two ways to like, let's say this to extremes. Right. There's extreme of do everything yourself and there's extreme. If just use blindly, something that someone does for you. And in that space of, you know, all the grad students in the world in Michigan, earning they spent considerable amount of time configuring their environments. That's a skill with love during PHD's and Docker and these things if you, if you don't become an IT guy or a DevOps guy, but just learn from the best there, and they do some of the things that on security, and that's really important for data that we have that I don't know. I don't have an inkling, but they expose us to AWS services exposes to some Docker stuff. So I'm not an expert Dr expert Kubis expert. But knowing a little bit of that. And they both empowers you to try more bold research ideas, and actually debugged. And when you care about the performance of your model, not just in terms of its accuracy, but it speed having. These knowledge Nabil's you to do research, much faster actually, which is counterintuitive a little bit. But again when you're beyond this, that's what it takes. Right. Right. You started out doing a lot of this yourself yourself, meaning within, you know, as research community research, scientists, it sounds like you are presenting with an infrastructure persons. You've got kind of professional support. Yeah. We do. We do really tightly with them. I also team is like probably like thirty percent engineers. Okay. And it's, it's really I think it's really good for research teams to have this mix of really scientists engineers, and because again, as I said, did lines are blurred at large scale research, and you need these skills, and obviously, also like the all the DevOps and infrastructure engineering team. So the collaborative spirit that you're is really, really good. Like because we're small we're very tightly knit. And because there was no technical debt. We're billing everything together and. Really nothing that the infrastructure injuring builds was done in isolation without consulting us. So that's why we have a system that works really smoothly because all the concerns were shared and address at the same time from all the pieces of the puzzle. So it's really nice to have that, like kick ass modern infrastructure built around around you somehow and with you. And so did that did that infrastructure engineering team. And support was that always there or did that come at a certain point after you'd build some things, it's a fairly recent addition. So we started kind of organically, and then you had some people that were there, and it started to be formalized only recently as we scaled up and where that became much more obvious. And is that infrastructure team primarily responsible for like where's the line that they how far up the sector they go? They worrying about like tools and frame. And software platforms. Or is it primarily, you know, infrastructure and, you know, network and does skin foul systems and connections to the cloud and all of that stuff. So I would say to that her. So I think the lines are blurry, but you need this single responsive with your principal. You know, plies well for software also plays for, you know, there's this Conway's law that says that. Software organization, right software. That's architect in a way that reflects the organization. Right. And so I think it's really good. If you have like clear responsibilities, but also the lines are a bit blurred because that means that you get a system that is flexible. But you need kind of responsibilities to so there's some separation and might team in Michigan research. And we are the ones that made the decision to switch to bite or trance and the way we did that is that for inspiring implemented yellow myself a year and a half ago, and all the difference deep learning frameworks. And it was after doing that, like section is really nice. Because it's structured prediction problem. That's shoehorned into a classification one. And so it breaks the that most frameworks support like from the get-go, and so if you use that, you know, you're stretching limited capabilities of the network in terms of the framework, API and reimplementing yolo in all the different frameworks made it clear. That as a research scientist, I value flexibility. And had to flex with each other is also very good. There's other alternates, but debugging stuff. So at certain levels, like that's why said, like research, scientists were making engineering decisions because choosing George is something that we wanted to make as a research scientist group. And for reason, also, if the particular research were doing so, for instance one of the things we're doing is the paper recently SuperNet, which is a paper about predicting the debt from of seen from a single image. And, and so he's self supervised method where he's geometry supervision instead of using labels because for that you can't label. And, and this is exemplary super resolution, so this idea of high resolution is actually important also for accuracy few super resulted images. This helps you predict better maps, one of the key findings that we made in in the paper. And so all that is also enabled because of the choices we made on the software. Sites and torch and all these kind of things, and also they're under community around it, so that Nabil's us to really move fast and sit on the shoulder of joins. So I talked to different organizations that have differing opinions on while how opinionated to be for their organizations. It sounds like you're of the mind to kind of standardize on in this case on torch at TRI as opposed to other places. We're going to build a kind of a framework platform, and it's going to be able to support whatever the research scientist or engineer wants to use talk me through a little bit of the, the way you think about that. I, I think about it almost mathematical terms, the by as in straight of, and if you have small bias right, and if you have a high vari-, you're really favoring exploration for these kind of stuff you need. A lot of people are willing to support you. Right. So if you say, oh, yeah. Slim, and Coburn, teas and by intensive so and everything end a little framework, that, that random guy made on his own free time. Then, you're, you're so first of all, like what is actually your business like, like, is it making those that infrastructure? And no for us. It's not for us. It's making some robots awesome missionary thing. So I clearly air more in device area, but, you know of. Map reduce right. Exploration expeditions, right off. Would you. I you have high variance and for a little while you go out you explore, and you're maybe not bounds by you implement yellow and every framework. Exactly, that's right. And then but then that's on point to make a decision. Right. That's not sustainable. And so, and you want to move fast, and it clearly identified direction. Once you have identified that direction and you never have enough data to prove that you're right. So at some point, you have to have express leadership and just go with it, and then you go for it. And of course, you keep an open mind, because then there's the next phase of explosion, because your rights for only short amount of time in, in this field of deep learning do, we take a diversion on the kind of the path that you laid out in the first kind of turn it step one, we got beautifully sidetrack wonderful direction. So, so. Yeah. So we were single nodes everything in Iran, and then, moved to like dry, existing storage solutions then move to more distributed file system. And once we had this. Because it's an in memory distributed, file system. We didn't have starvation anymore. But then our training was slow because we were to single machine and then Peter instances happens, we start to use the hundreds of us much better that regardless the tuning storage again to avoid P starvation, and then we again admitted to multi mode, and with the district fell system that at least the data was easily accessible from all the different notes. And then, that's when we start to hit limitations of like this related pie torch which was very recent at the time before we jump to distribute it. I'm curious about the you've got some. I guess, quote unquote, high-performers, like virtual CPU's the machine configuration parameters like you know, they're kind of universal rules of thumb for that kind of thing that you figured out, or do you experiment with it a lot is that job dependent overly focused on economic optimization like had to work through all that. So we up team is for time. We don't have to miss for the okay. That was easy. That was easy. We haven't. So that's more job at the infrastructure during people's does that mean you just get the biggest one with the best DP and got it. That's exactly it, so, and also because our workloads was just that was the only thing to do so go big or go home. That's basically what would yeah. Yeah. So for single machine just tried to scale as much as possible on single machine that meant big, big instances, we've site, too soon d then you wants to announce the even bigger actually that's feedback that we directly gave AWS. It's quite cool to see that, that we give them feedback a year ago. And then, like keynotes was, oh, and we heard you. We did this the biggest instances that they made. That's, that's something that we had asked for an a couple of other stuff. So but you're still on a single machine. And so when you are at kind of topping out at a single machine. How long were you jobs running for? So at this age it was more in the order of weeks. That's kind of. Job is so, so the main one in terms of competition, the most expensive when Simmons expectation because again it's like high resolution it's very dense it's prediction. And so that, that was the most competitions expensive job, another type of job that we do that is also very expensive is imitation earning so we do a lot of research on two driving. The main reason is not so much that we believe that it's all you need to driving this Leonardo. But we get a lot of data from actual cars, and, and so we get a lot of demonstrations. So there's this really interesting research that we're working on, which is homage value, derived from these demonstrations, this form of supervi- supervision, driving that, you want to still down into your models. And so we do a lot of research there. And that's. Use all the data is really the question that animates. How can we use all the data because we can't label everything we're not going to active learning routes and the same thing that everybody else is doing because we're doing that. But that's not the open research challenge. Everybody knows active learning is a good thing to win labels things. We're really interested in self supervised learning. How can we really use all the data by leveraging geometry, for instance, how do we use demonstrations at scale? And so that's those are the workflows because motivated by the research direction. We're going in those were the most intensive once and single machine these are things that easily take weeks. Okay. So then that necessitated jumping over the distributed training. Yes. Absolutely. Did you do that after the decision to go with pine torch? Or did you have to figure that out twice? No, we had made. So because also, we have a lot of like wearing Silicon Valley. So it's really, it's really nice. That's there's a lot of dense communique..
"research scientist" Discussed on Thunder Radio
"Back to this hour of the doctor Bob Martin show, our special guests. This segment of the program is rob Martin a nutrition scientists longevity research, or we're talking about telomeres those capsules all important little caps at the end of the chromosomes real literally the holy grail too long. And our goal in robs goal is to help. Teach you and give you the the means by which to keep your telomeres long and strong as possible because in doing so not only will you reduce your risk of future degenerative, mostly preventable diseases. But also if they're on the burn too fast right now, it could be the cause of all the symptomology that you have your body talking back to you with all those idiot lights going off in your body telling you that something's wrong, and then you're not getting results anywhere else. So before we took our break, their, rob. I asked you to be prepared to tell people because I know you've spike the interest of many by talking about this topic, which is not a common topic that people hear about they certainly not are hearing about doctors offices. I can tell you that. How did they get a hold of the cello vite supplement that the human studies show that are L thing to lengthen those all important organisms? Absolutely. And the Gulf between the practicing physician, God bless them. And the research scientist is like the Grand Canyon. Okay. I live in the realm of what the research scientists live. Okay. So here we go. We have the best offer ever for the televised for just a dollar a day. You can start making your cells younger just like those middle aged and senior men and women.
"research scientist" Discussed on WDRC
"And people are up in arms at the cost of his treatment at 850000 and i'm pulling my hair out i'm watching all of these people yell and scream and how they charge so much how much do you think it cost 5th develop this type of therapy now these type of blindness only hits about 2000 people in the country how much money do you think it cost to develop this philly you think that was free to you think that was free you go all of the all of the doctors all the research scientist everything that went into this not cost anything again people we need to figure out some way we start teaching people how younger level in schools how the world works watchdogging wall streetcom watchdog on wall streetcom get there pre consultation right free reign we're not charging for the consultation dollars but it's not free in the sense it costs us money because it cost us time so it's not free watchdog wall streetcom get their take advantage of all the great stuff we have we'll be back bruce more calcium is the watchdog the bottles yeah.
"research scientist" Discussed on AP News
"Leaves largely spared by gypsy moth caterpillar look healthy full each experts say that all suggest an optimal season minke magazine's annual forecast predicts the particularly strong and vibrant display thousands of people from all over the country of made their way to parts of the country in the path of totality for monday's solar eclipse masol lunar research scientist noah petrol says the timing of the eclipse will depend on where you are she but son get completely obscured and what you'll bentley is a beautiful solar corona become apparent folks outside the path the totality would see a partial eclipse which is still spectacular to say he says the eclipse will have a number of noticeable effects even for those outside the path of totality matter where you are along the path it to tell your after the path the totality you'll see something different you'll fuel something different temperaturewise and you actually hear something different where animals are we'll start making their nighttime noise and petrol says the moon's top og review will produce some irregularlyshaped a shadow on the earth actually a forty nine cited polyglot in the case of eclipse this year and that's because each valley along the side of the moon edge of the moon actors a little projector so you can imagine fortynine little projected um son sunrays onto the earth handy says no matter what makes sure you're wearing proper protection when viewing the eclipse to avoid damage to your eyes this will be the first total solar eclipse a ninety nine years to cross a coast to coast swath of the united states for the ap rubble of firms with wherever it it take a closer look at unlimited data plans you'll see they're not always up front.
"research scientist" Discussed on WLOB
"Is trump like to threaten north korea is it longoverdue or is it a bellicosity that the world doesn't need right now it's as simple as that and i wouldn't say america's divided on this i mean my gut instinct tells me most americans would back trump if he was to lawn if he will launch some kind of strike against north korea preemptively or otherwise taking out command and control centers taking out nuclear facilities taking out there the research scientist barracks anything along those lines america which year and by the way so what japan incidentally lee south korea would philippines would the only people would oppose it would be jerry brown and the maxine waters because they need to understand i mean they want us to understand that they need quietude around their own uh deficiencies here's maxine waters saying we have to find that if we can work with north korea what do you mean we have to find out we've been trying to work with north korea since since the the armistice maxine what world are you living in and don't you love how she's calling for diplomacy all of a sudden that big mouth liar after months and months and months of shooting of a filthy mouth screaming for impeachment day after day after data hud dump constituents who she rips off she lives in a multi milliondollar house outside our own districts and here's a woman who stirs up her own constituents would hatred for the president day after day and she's now screaming for diplomacy this is astounding to make but i'm asking you is trump right to threaten north korea you know there was a secret police chief in the soviet union they move beria one of the most frightened fearful people on earth and he had a statement it was show me the man and i'll show you the crime mueller is now the barrier of america show me the man and i'll show you the crime see breaks in the doors of mr manafort the day after metaphor testifies by the way to cooperates with the senate and the fbi the fbi uses soviet style tactics of kicking in his door with a fake search warrant gotten for some crackpot judge and now we're living in the soviet union and what was sitting in not recognize using what's going on in this country from the american socalled democratic left xiaomi demand and i'll show you the crime but.
"research scientist" Discussed on 1410 WDOV
"Figured out um and i'll tell you again you had richard hoagland a profound space research scientist on your shokhin in into incidents is he reported that really got my attention again i've been in the business of of detection with physicists and i understand how we make detections none of fast a vast level in my i think it would call depeg assists incident member the pegasus incident i remember it very remember the giant disc that showed up on radar show up in the visible spectrum yes that's exactly what i'm looking for you're looking for a high energy signal of something that isn't appearing in lower dimension why don't think there's a big argument i i all those were two to three hundred mile white circles and they will moving and they showed up on radar by no who didn't see them that is exactly what we should be looking for in regards to uh considering einstein lauded a solid craft can't do the speed of white could if you turn into pierre energy you can that's what we should be looking for how does that happen how do we how do we take math and turn it into light and how is it possible that these uh that that for eight app light years away these are crafter able to break the journey and much less time than that well that are subject it in in the video in the documentary and also in the book i treat with the incredible uh respect on what i did is i would with a zero point energy scientists named steve oakland in maui we got a giant tv flood.
"research scientist" Discussed on The Limit Does Not Exist
"The next time you wait even longer interesting death that's certainly explained kate why you can't remember anything that you cramped brain yup you are a thing out that learning yet so okay super before you join dueling and go you you order phd in computer science from the university of wisconsin and then worked as a post stock researcher at carnegie mellon so what's it like to be a research scientist at a university verses at a tech company there well a lot of the problems with the same actually and a lot of the things that i've learned in an economic setting i apply every day and startup kinda industry setting but i think one of the biggest differences what i would call a the reward function for a job well done and in an academic setting the reward function this is something like the number of papers that you publisher of the number of citations that gets that's less of an emphasis in eta especially a startup uh where uh the reward function is building a product that does a better job at acquiring and retaining its users in in our case and and in also teaching them better yeah that's so interesting 'cause you're talking about like a number of different papers verses sounds like continuing to add value and expand this one distinct product is that race right so will i mean the daytoday work is often pretty similar we have lots of data we have some hypotheses we build some computational models we test them to see if they explain the data well and then uh rather than going in writing a paper on it most of the time the next step for me is to implement the idea in the production system more.
"research scientist" Discussed on Science Friday
"Susan thompson keppler research scientist at the said he institute and lead author of the new research joins me from the campus of nasa aims in silicon valley welcome to the program hi it's the so how many planets is this now uh we are up to just over four thousand planet candidates once you had this catalogue in the we've only confirmed about two thousand thirty or so of them how and released ten that could be rocky planet yeah we of with this catalogue we able to extend out and find really long period planets that are in the habitable zone and really small planets that could be rocky there about the same size as the earth using the capspace craft let's let's refresh our memory forest would you please were about how kepler finds the planet scooters so far away so keppler is like a big bucket for light were just measuring how bright the stars are and what we're looking for is the planet passing in front of the star and that causes the started to suddenly look a little bit dimmer for just an hour to or may be up to 12 hours and um we see that happened several times and once we see that happened several times we call it a planet candidates and that's what we were collecting in this catalogue absorb these two hundred planets are they all from the same part of the sky or the you'll spread out all over yet this was all from kepler's original mission where we're looking at the part of the sky near sickness the swan so you have four thousand candidates in total from this one area.
"research scientist" Discussed on How I Built This
"You know it hardwood floors like sort of distressed would floors and beautiful appliances and a bright yeah so we were more conservative back then in terms of our investment you know we did all lot of the work ourself immune to delayed the actual labour yet there is a limit thomas they would build verses us and so we had some responsibility one of our responsibilities was id wearing all the cabling and so we were out there getting bids for it wearing our member we got one for a hundred thousand dollars one for like eighty seven thousand five hundred i remember getting those estimates mean like hundred thousand dollars a pretty round number for wiring lake each and you have like a more detailed calculation of how much that cost so my little brother and i he's background in he was premed and he's the kind of research scientist as in his brain and he had moved out to help us it was nikulin adam at the time and unlike hey why don't we look into how this actually works and duly his brain works is then to break it down into like the most micro parts basically every cable ran how long as it how many terminations you know how do you do terminations then i mean i remember he estimated it's going to cost us the eight thousand dollars in materials and then you know whatever like fifty hours of labour whatever the number was it was some crazy small number relative to this hundred dollar price and so were like well it's not rocket science and we can do that are self and so we just did ourself wait you flyer and the building yourselves yet you're like lane the wire yeah our and all of it so we had it like a hammer drill and like drill into the brick to hold like brackets to hold the wire we had to amino so much fun it was like it was crazy but it was also lake were really fully engage and we had to teach ourselves how to connect patch panels and program switches in you know so we put in that whole infrastructure am i remember when things we had do is actually um it's this thing called soda blasting which is if you want that kulik exposed brick gan it's been painted have to blast the san asked the ena this been called soda blasting were used baking soda do that.
"research scientist" Discussed on KFI AM 640
"Am 640 james mahaffy has a bachelor of science and physics a master of science and a phd in nuclear engineering he is a senior research scientist at the georgia tech research institute has worked on projects for the defense nuclear agency the georgia power company the national ground intelligence center he's got the background and here we are on coast to coast talking about his work atomic adventurers james welcome to the program i adored i'm glad to be here i'm looking forward to this i think you're going to educate a lot of us about the atomic world much more than we probably have any idea what's going on and how would exist but how did you get involved in this particular field james our george attack got a degree and civic life in seventy three and a very very with a bad ear for high of no the apollo space program it to shut down and there were rumors of the the delivering pizzas and so on who i think they do go to grab who oh i was very interested in computers computer effective were different from what they are now and those who knew jerry i had a dual degree he hit a masters and new cringes nearing and then a separate the degree and computer engineer dario trump to go out and graduated with a master uh and uh it new printed airing but there were no job new grant an area so really i would have thought it would you'd you'd flood the market well uh in the nineteen 70s knows when nuclear power dried up oh they stop building power plants are already had enough of them and uh it would dead the thing from them i really so i want you to do is go back and get another degree time i was working at engineering burma station georgia terek which eventually came greet here are the tech researcher i do at i worked on secret project for the parliament have found a very interesting work and uh ipod i graduated with.
"research scientist" Discussed on The Pulse
"Counter is a longtime businessman she partner with the research scientist create jump pro judo window startups like this are working on therapies for cancer alzheimer's autoimmune diseases and more but conner says the competition for early funding is fierce even if you really have a really promising concept with a proven technology you have to get people to believe in it and to believe in you years ago conner might have had to convince big pharma to believe in him but today lots of the big pharmaceutical companies no longer develop drugs from scratch instead there looking to pick up products that have already advance them and are showing more promise this void plus an explosion in hope biotechnology field outside that traditional pipeline has created a new marketplace for early stage investment it is fragmented and a little chaotic swat the chaturvedi is founder of prophylaxis it's a company in the bay area that's trying to held by connecting angel investors with biotech startups so the number one challenges for these companies how do they access these individuals rant vice versa for the individuals how do they access these companies that she says getting investors interested in this area is tricky science is intimidating and complicated most people don't understand them green bay's complex technologies so typically what has happened is that has for the natural the pool of in jilin lestas and she says this biotech and pharma area runs counter to the general startup culture that's driven in large part by the software an internet world where success is often built around celebrities pr and buzz so for example how many people are downloading some app makes them more popular than valuable so more investors want in but an signs that is not the case in science would you should be asking for is data what is the experiments that you've done what will the results of what were the stress destituted how long did you do them so those kinds of things are more important but is this really the best way to develop new healthcare technologies and therapies so there are pros and eric hines rachel sachs is a professor at washington university in st louis xiii studies this whole innovation and biotech area.