35 Burst results for "Carnegie Mellon"
Gold just hit $2,000 an ounce — but that could be a scary sign for the economy
"The price of an ounce of gold is now up to $2000 but analysts say this could be in people are losing confidence in the stock market. Carnegie Mellon Finance professor says people are buying gold during the pandemic has a safe asset, but the price is not
Using Your Brain Without Thinking
"What does it mean to use your brain? And how is that different than just thinking? As developers engage in thinking all the time but here's a entirely separate part of our brains that we might be missing out on using. That could be better at solving some of the problems that we face on a day-to-day basis. My Name is Jonathan trailer listening to develop for T and my goal on the show helped driven developers like you find clarity perspective and purpose in their careers. One of the amazing things about the. Human. Brain. Is Its ability to process complex topics. This is why we can write code that is abstracted so many levels. Away, from a physical reality that we have to tangibly think about. We can imagine entire. Kind of universes where we can create stories and. keep track of those stories while we read a book. A book that was written with a bunch of characters that are enough themselves abstractions. These are characters that we may not have ever even seen that specific character that specific size before. But somehow we are able to process all of this information and create meaning out of it. This is an incredible feat and part of our kind of intellectual superiority that we are aware of the domination that we have over the world around us. Has Given us. A somewhat distorted picture of what the brain is actually capable of more importantly where the limits are. And it's very simple to see the limits of your brain and specifically limits that we're gonNA talk about today. If you want to test these limits you can. Try to brute force memorize the first twenty digits of Pi. This isn't a lot of information. It's just twenty digits in after all we can process a lot more. Information than just twenty digits, we can read entire books with thousands of pages and understand them. So what is it about remembering twenty digits? Makes it difficult? Here's another exercising might want to try. that. You've probably faced already in your career, go and look at the features of what say three or four different libraries, popular libraries or three or four different languages and try to decide which one is best. This kind of information that you have to process. It's really difficult to do because the number of variables and that's the critical factor for today's episode, the number of variables that you have to weigh against each other. Can Be really large temper variables. You can imagine for example. That you're trying to deduce which which language should you learn next let's say you're a beginner programmer and maybe you're trying to decide which language to learn. You can use variables like the market size. You can try to quantify how much you enjoy that language or. Even how much you expect to enjoy it in the future, you can imagine you would use measures like the number of available repositories on get hub or get hubs own report of the trends for a given language. How do you decide what trend to use or how far back to look? These are all different questions they you would have to try to answer and then compare between the different languages. And so now you have this very large list of pros and cons and. You sit down and try to look over that information, but this is. Where we hit our limit. Our ability to cognitively process or think about something on purpose. We only have so much capacity to think in parallel. This is critical factor remember again, the number of variables were very good about thinking about one thing. At a time. In fact, most of the advice that you receive on this podcast is an attempt to get you to think about fewer things at any given point in time and reduce the things that you are working on to the simplest form. So you don't have to keep a lot of information in your head. But if you are trying to make a decision complex decision with a lot of variables. There is another part of our brains we can tap into what's interesting is that as knowledge workers, we are paid for using this one specific part of our brain, this prefrontal CORTEX. The part that's responsible for thinking very deeply and thinking very focused manner. But. There's another part of our brains that can help us think more abstractly. And without the same limits of the cognitive processing limits, the would find in the prefrontal CORTEX. Lots of studies. For example, one from Carnegie Mellon support the idea that the rest of our brain is working on the problem. In parallel to us focusing on other things. For example. If you expose yourself to all of the information about the various programming languages that you're considering let's say you have four of them. Then you can go and do something totally unrelated to that. Your going to keep on working on that decision problem. Now, we're not really consciously aware of this and there's no way to become aware of it but once we return to that problem at a later point in time we may have a different sense of clarity and we might even have. We might feel is a gut intuition, but actually it's an intuition that was given to us by that unconscious processing that's happening in the rest of our brain. So. Here's the critical thing to to take away I. We said the the most critical thing is to remember that this has to do with the number of variable. So if you can reduce the number of variables that you're thinking about, then you can actually process those entirely in that prefrontal. CORTEX. For example, if you're working on a math problem, this is a perfect example of processing in the prefrontal. CORTEX. But if you're working on something that requires much more evaluation much further a can of discussion about multiple variables or a comparison between multiple things, and that's not something that you're going to be able to hold in your prefrontal Cortex, the working memory for of a better explanations too small. So the prescription to fix this problem is to expose yourself to the information all the relevant information for making a given decision and then go and do something else. Maybe take a walk give yourself something that's totally unrelated that won't allow your mind drift backing and try to process that information again, on purpose in that intentional and conscious way.
"carnegie mellon" Discussed on Scientific Sense
"Welcome to the site of accents podcast. Where we.
How to Become a Change Agent in Your Health System with Tony Manuel
"Welcome back to the outcomes, rockets, Sal Marquez is here today I have the privilege of hosting Dr Tony Manual Dr Tony. Manual is a practicing anesthesiologist and Austin Texas. He's a partner with the United States anesthesia partners central Texas and has been in practice since two thousand two. He's an assistant professor. In the Department of surgery and Peri, operative care at Bell Medical School Dr Manual received his undergraduate degree from Vanderbilt attended the University of Texas Health Science Center for medical. School, completed his residency in anesthesia at the University of North Carolina or Or. He was recognized as the outstanding resident and fellow cardiovascular anesthesia at Duke, university in two thousand seventeen. He received his masters in medical management degree from Carnegie Mellon University and today he's playing. Multiple Roles as as he has in in his career and today we're going to be talking about physician innovation, and in particular how physicians can evolve their career to be greater contributors beyond the point of care and so. I WanNa thank you Tony for joining me today to have this very interesting discussion with you saw thanks so much great. Great to be owner podcasts, and I WANNA. Thank you for actually doing this podcast. Because for a lot of people like myself, it's been a great conduit to learn about what other people are doing, and what best practices that are out there, and it's an alternative to sort of the Journal Review articles that we have historically read and I've actually looked up several companies that you've had on and engaged with them. Really appreciate what you're doing I. Love that man now. That's great. I'm glad to hear that you've done that. That's the intention. Intention and so I appreciate you for doing that, so you know we are having a discussion. Folks Tony and I connected and said you know what the role of the physician is changing, and and so what I wanNA. Do today is just highlight how that's changing through the life of Dr Manual here and so you know I love for you Tony at to just kind of walk us through some of the work that you're doing and how it's changed from just practicing to actually doing more You know as we engage this. You know three five trillion dollar industry that call healthcare. Yeah, it's it's been an interesting journey for me and you know have to credit one of my anesthesia attending when I was in residency, his name's Dave mayor said Gimme, grapevine goes Tony You have to continually strives to maximize your career and Let's see well. What does that mean well? You definitely want to start trying to be the best clinician you can be once. You achieve that you should really look at you. Know becoming really strong in other areas, and I always took that to heart in so I think back to when I first started here in Austin I became the division chief of cardiovascular. Cardiovascular Anesthesia Rochester, saying I helped create division of cardiovascular. Because at that time we were Basically, everybody was doing it, and I fell coming out of myself. This'll be really better if we limited number of people at work in that space and you know put together some protocols and got the team together, and we saw some really good outcomes from that work and I fast forward to what I'm doing today, and that work has changed so much partly because I think every clinician you have to get educated, and I use a rudimentary tools back then, but in after getting that masters degree from Carnegie Mellon I really developed at toolkit that allows me to take on. On much more complex problems that we face and healthcare today. Yeah, that's so interesting, and so you have that entrepreneurial bug from the beginning right so you kind of re retooled the way that you guys approached cardiovascular anesthesia and I'm sure with with much improvements and outcomes, but then you've taken other steps to. You've been involved in startups, and now you're doing different roles. Can you talk to us a little bit more about that? Yeah I worked my way of the medical staff leadership and ultimately became the president of medical staffing while that was a great experience after I graduated with my master's degree. The entrepreneur apart really was intriguing to me. In more important is the. The interface between the clinicians in technology and so The startup is called Dynamic Lights based here in Austin, and it's actually technology out of the University of Texas and They had great idea concept. Basically, it's how to noninvasive map blood flow during Sri will hand you an aneurysm surgery and uses what's called speckled laser technology and I was like honestly Craig. It's continuous. It's noninvasive, but they never really thought about the interaction. How you get it. It's dockers hands. How do you test it? And that was sort of my strength and so I, said well. Let's work together and figure this out and to date. You know we've incorporated. We're FDA approved, and we're. Ducking clinical trial and we're looking to partner with a couple of larger health tech firms,
Ivy League suspends fall sports due to coronavirus pandemic
"League schools announced today they will not have any sports in the upcoming months. The eight prestigious Ivy League schools have cancelled all fall sports because of the covert 19 pandemic. A decision on winter sports is expected later this month. Carnegie Mellon University in Pittsburgh is also canceled All fall SPORTS The 11 members of The Centennial conference that compete in the Division Three have also cancelled all fall SPORTS because of Corona virus.
Rethinking Model Size: Train Large, Then Compress with Joseph Gonzalez
"Art. Everyone I am on the line. With Joey. Gonzalez. Joey is an assistant professor at UC Berkeley in the e S. Department. Joey welcome to the PODCAST. Thank you for having me. I'm really looking forward to diving into this. This conversation and in particular talking about M. L. Systems and your recent paper on train large then compress. But before we do that. Please share a little bit about your background and how you came to work in. Yeah excellent so my stories of it. Funny I started my PhD at Carnegie Mellon with an interest in actually flipping helicopters. Because that was the thing to do back in two thousand six awhile back. Lipping HELICOPTER LIVING HELICOPTER. Flying THEM FIXING THEM UP. Sell or fly them and then flip them. Actually a colleague of Mine Peter Beale now at Berkeley When he was finishing up his his thesis work is looking at how to do interesting. Control for helicopters cool and I knew I was. I went to my thesis advising the you've worked on control as well. I'm kind of interested in flipping helicopter. I think that's that's really neat research and you know that was And it actually was some of the pioneering work that we see today in reinforcement learning. But what's kind of cool about? The story is my adviser at that time being a real machining researcher. I was like you know what flipping helicopters. That's that's that's exciting but there's something more important like we can actually help the world with sensors we can build sensor networks to to monitor fires. We can use principle machine learning techniques. I should add that when I was looking at the flipping helicopters we should flip them with neural networks. And the other thing. My advisors said which was good advice at the time was a neural networks. Really serious research. We use more statistical methods graphical models things that have formal foundations that we can reason about right kind of detailed analysis and understand what our models are doing and that was good advice and so I went down this path of how to build a proxies Beijing per metric methods to reason about link quality incident works and in that process of doing that. I kind of stumbled into it problem. I was reading a lot of Matlab code to compute big Matrix in verses and then approximations that to make it run faster and one of the things I enjoy doing in the process of exploring these efficient matlab programs was trying to make them more parallel and I think advisor clued in is a good is like you know what maybe you enjoy that more so maybe instead of focusing on the per metrics and the sense now works. Let's start to think about how to make machine learning more efficient in particular at that point in time Duke was taking offense. Map Produce. That's GonNa Change machine learning and we were thinking well we're working on graphs and they just don't fit the map produce pattern and the kinds of computation. We were doing just wasn't it didn't actually fit the technology that people were building so we started to explore a different design of system so designer systems for computation graphs which took down the design of of graph processing systems system. That I ended up writing. His kind of the end of my thesis was a graph lab for doing very large analysis of graphs and so by the time I finished my PhD. I was actually writing systems papers. Not machining papers in the field was changing very very rapidly to this around two thousand twelve and if anyone's been following the history of machine learning around twenty twelve everyone started to realize maybe actually neural nets for a good idea. The deep learning these ideas actually really dated back to Nineteen Eighty S. They're actually really starting to work a and they were changing. The field machine learning and grass are also taking off so we built actually a company around the systems that I was developing as a graduate student who was graph. Flab that evolved into a company for building tools for data scientists do interesting machine learning at scale that was ultimately acquired by Apple. And around that time I also joined the UC Berkeley employed as a post doc. A chance to come out to California and it was really exciting opportunity to do research in a different system. A system called spark which eventually became Apache spark and there we started develop the graph processing foundation for the Apache spark system and again as I started to explore more and more into the field I learn more about research data systems and transaction processing and how those connect back to machine learning and so after finishing my post doc I came to Berkeley in fact I chose not to follow the much more lucrative path of the and it was going to ask about that. A made a terrible financial decision. But I'm happy because I have a chance to work with students. I'm a little less happy because I'm not as wealthy as one could have been but now I am teaching students that do research at the intersection of machine learning and systems and so we have a pretty broad agenda around how to better of technologies for delivering models to manage machinery. Life Cycle not just training but prediction how to prioritize Training Experiments on the cloud to use service computing to make machine learning more cost effective and easier to deploy. We have a big agenda around autonomous driving building the actual platform that supports autonomous driving not necessarily the models but how they are connected together to make a reliable car and we have work in natural language processing and computer vision and one of those papers when the hoping to talk a bit about today which is our work on on making Burt models easier to train and it to has kind of funny story how we came to To actually realization that what we were thinking was entirely wrong. And that's what that paper talk a bit about.
reCAPTCHA and Duolingo: Luis von Ahn
"Think about the small moments or decisions in your life that actually had a huge impact on how your life turned out. Maybe it was a conversation. You struck up with the person next to you on an airplane. Maybe it was a party. You reluctantly went to only to meet the person you'd eventually marry or maybe it was a decision to stay on vacation an extra day that sparked a new idea for Kevin System. It was a random remark from his girlfriend that made him decide to use filters on instagram for Blake. Majkowski was a chance meeting with a group of young Argentinian who took him to the countryside where he saw kids with no shoes. That one day inspired him to create. Tom's and for Louis Fun on it was a free lecture at Carnegie Mellon University in two thousand. We'll get deeper into the story in a few minutes but that single lecture would lead him to invent to ingenious new tools the I was capture. Yes captures those annoying twisted and blurred letters. You have to type into a website to prove your human and the second one was duo lingo now. The biggest language learning APP in the world which is now getting even more popular because people are looking for new things to do now that they're stuck at home but was captured and duo. Lingo were designed to harness the power of crowdsourcing to solve problems. And I'M GONNA blow your mind here if you have ever typed in a capture or reused dueling go. There's a good chance you've taken part in a massive online collaboration that you probably weren't even aware of and it's amazing. How Louis came up with all this but let's start at the beginning. Lewis was born in Guatemala in late. Nineteen Seventy S. Both as parents were doctors and though he was surrounded by poverty violence in Guatemala City. Louis screw up in comparative privilege and as a kid. He spent a lot of time hanging out at the family business. My Mother's family actually had a candy. Factory everybody is always a Mesa. The fact that I grew up with a candy factory they think it was like Willy. Wonka or something. I was not all that much into the candidate. Self I was into the machines because basically the candies made by these gigantic machines. That bump out I don't know how many thousands of pieces of candy per hour and basically all my weekends. I spent playing at the Candy Factory and I would They the machines apart and put them back together they would be some extra pieces after. I put him back together on that. That would be a problem but what? What kind of student were you were? You were school pretty easy for you. Yeah I was pretty nerdy basically. That was really good at math. Math was just easy to me. I what I would do during the summers is basically get either next year or you know. Couple YEARS LATER. Math books on all the sizes. Wow it kind of came easy but the way I really got good ideas by doing hundreds and hundreds exercises. That's what you do in. The summertime was bored. I mean I was an only child I is. I didn't have that much to do. This is remember this is also pre Internet pre everything. So what was I going to do? Man That's what I did was putting playing cards in the spokes of my bicycle and by jolly ranchers seven. Eleven should math books. So you were. Did you just love math? I mean it sounds like kids. Don't think about their future. They're not like I'm going to study math so I can be in tech one day like unless I've really enjoyed it. I I enjoyed it was it was like a puzzle for me by the way this is not the only thing I did. I mean I I also played a lot of video games Pirated Video Games in my commodore sixty four like floppy disks. Floppy Disk loppy discs. That's right I wanted a Nintendo. When I was eight my mother would not get many intendo. She instead got me computer. Commodore Sixty Four. And I couldn't figure out how to use it but eventually I read like the manual stuff and I figured out how to use it more than I figured out. I could buy other people's video games. And so I became a little hub in my in my little neighbourhood but these were not other kids adults or kind of basically young adults who had a computer and they would come to my house and I would take their games and give them my games exchange so then. I collected a pretty large number of video games but sh- mentioned right that I mean because your childhood sounds pretty nice but but like as a kid I guess or even as a teenager there was a civil war in Guatemala right. I mean we know that today. There's a a lot of violence there. Obviously violence in the US and other countries to but Guatemala's has been particularly hard hit. I mean did it feel dangerous when you're a kid yes it did. There was a civil war pretty much since I was born in seventy nine to nineteen ninety-six. There was a civil war going on the whole time. It always felt dangerous when I was fifteen or so. My aunt was kidnapped for ransom. I mean she was gone for seven or eight days. Wow People's cars would be stolen. I don't every couple of months. Somebody's car would be stolen in my family. Going past seven thirty PM was rare games. You needed to go out in a large group. If you're going to go up at seven thirty PM and I did my house had walls and barbed wire yeah. It felt dangerous. I mean this is one of just one of the reasons I came to the US. Actually I mean I was. After my aunt was kidnapped I thought to myself. I don't WanNa live here. Yeah and I guess you did end up leaving Guatemala for college because you went to Duke in North Carolina and you describe yourself as a like a math nerd in school and and is that what you intended to do like to do something in math. That's what I wanted to become an economic math professor. I was pretty certain. I wanted to become a math professor at the time. I thought the best thing that I can do is really learn a lot of math and I really it and I thought it was futile to learn how to deal with other people. It is interesting because my job. These days is one hundred percent just dealing with other people's problems. I'm just trying to understand the so so by becoming math professor. You thought. Hey I wouldn't have to deal with people I would just deal with facts. Data and numbers. Yes yes and you know I. I'll do math research all day long. And every now and then after class of but whatever that's like a tax That's that's what I thought so all right so you are She gets your degree and you this path to go into academia and you go into a PhD program at Carnegie Mellon Correct and I guess you go into computer science right yes. I changed from math computer science because I visited a math Grad school and what people were saying the professor was saying. Oh I'm working on this open problem that nobody's been able to solve for the last three hundred years and I thought I don't think I'm smart enough if you haven't done it and nobody's done it in three hundred years that's Kinda not for me whereas when you visit in computer science I mean this is crazy thing before like. Oh Yeah I still have an open program yesterday. Well it's a much younger field yet so that I thought that was much more exciting for me. At
Felicity Huffman's daughter Sophia accepted into prestigious university 1 year after college admissions scandal
"Says and Felicity Huffman's daughter is being accepted into a top university Sofia Macy is heading to Carnegie Mellon university's theatre program after mom's rolled in last year's college admissions scandal her mom pleaded guilty to paying fifteen thousand dollars to alter Sylvia's SAT answers well Sophia took the test again on our own got accepted into the
Facebook map shows you where people are reporting coronavirus symptoms
"Let's talk a little bit more about another couple of other efforts actually multiple efforts to track. Where infections might go next by having people self report systems symptoms rather facebook partnered with Carnegie Mellon University's Delfi Research Center on a survey asking users to report their symptoms both Carnegie Mellon and Facebook have now published websites with their initial. Findings can see a little heat map Go TO COVA CAST. Cmu DOT EDU. And you'll be able to see these heat maps later this week and eventually provide forecasts based on the data to help. Local health officials anticipate where hospital capacity needs might spike next facebook's own site provides a symptom map of the US at the county level showing what percentage of the population has reported systems county by county facebook's also partnering with the University of Maryland to take that survey global and Carnegie Mellon is building an API so let researchers access the data.
Facebook launches map to help identify coronavirus hot spots early
"In the social media giant is opening up about how it's helping fight coronavirus founder and CEO mark Zuckerberg penned an opinion editorial for The Washington Post in which he offered up the vast resources of the social media giant to help fight the covert nineteen pandemic his reasoning is straightforward accurate county by county data is needed from across the U. S. and Facebook has a community of billions of users globally and as we have all learned over the past few years I spoke knows a lot about us Facebook has already rolled out and often a symptom survey run by health researchers at Carnegie Mellon University and Facebook says answers to the survey were sent to the researchers not kept by the social media site so far Carnegie Mellon says they're getting one million responses a week in the United States Zuckerberg says the social media data is a new super power in fighting pandemics and urges people to use the data
The Evolution of ML and Furry Little Animals
"You are listening to talking machines Catherine Gorman Lawrence and Neil. We are again taping an episode in front of a live audience digitally recorded though on on talking machines. And if you want to be part of our live. Studio audience big quotes. You can follow us on twitter at Ti Okay. N. G. M. C. H. S. Or hit us up on the talking machines at gmail.com and our guest today for this interview on talking. Machines is Dr Terence. Annouce key doctors and thank you so much for taking the time to join us today. I really appreciate it Great to be here so we ask all of our guests the same question I. How did you get where you are? What's been your academic and industrial journey. You're also very involved in the reps conference. Tell US everything well. A wise man once told me that careers are only made retrospectively and I have no idea how he got here. There was no plan. It went through a sequence of stages starting with graduate school at Princeton in theoretical physics. From there when I finished that I for reasons that have to do with the field of physics. At the time which was a little bit more bummed I went into neuroscience so that was a post doc and then from there that's when I met. Geoffrey Hinton and had changed my life because we met him at a small seminar here in San Diego and set nineteen seventy nine. We hit it off and From that over the next few years you know blossoms the the Boehner Sheen and back prop and you know. The rest was history. Terry who you post talking with where you post talking in San Diego no no. This was a post doc at Harvard. Medical School in the Department of Neurobiology with Stephen Kofler who was widely considered to be the founder of modern neurobiology and It was an experimental post. Doc I actually recorded from neurons. Subic seventy nine. You mentioning physics. It was a little bit more bond a in some sort of connection modeling. That was also a very quiet period. That wasn't a lot going on it. Was this sort of age of classical. Ai Right you're absolutely right. This was in fact. It was the neural network winter. The seventies and it was primarily because of the failure of the perception. That's neat because you say failure of the percents on I read about that a lot. Do you really did fail. All was the men's ski paper little. What the mid ski books are in Minsk. Eighty books have killed it but was it a fair representation. Well you know it's interesting. I think that that's the myth that that book killed it but I actually think that there are other things going on and and Rosenblatt had died as well which seems pretty significant. Yes well He. He was a pioneer. But you have to understand that digital computers were regally primitive back. Then you know that even the most expensive you know the biggest computers you could buy. Don't have the power of your wristwatch today. Rosenblatt actually had to build an analog device. It a million dollars in today's dollars to build a analog device that had potentially otters driven by motors for the weight sums the learning. Wasn't it potentially because you know digital computers? Were good at logic but they were terrible. Doing a floating point is amazing so he built that at Cornell. Right that's right yeah Funded by the owner. Any case by by the time that we were getting started computers was the vaccine era. It was becoming possible. Do Simulations You know they were small-scale by today's standards but but really meant we could explorer in a way that Frank Rosenblatt couldn't so what you're saying around the perceptual and so just forbid of context for Central and sixty one. Is that right? It was fifty nine. I think it was the book but you know it was in that era of early sixty zero and so then there's this period where the digital computer actually wasn't powerful enough to do much and then digital kind of overtook and divinity but these analog machines would just now impractical from a point of view of expense. So you're saying it's less the book and more of a shift to the Digital Machine. That in those early days wasn't powerful enough to simulate the perception. Yes so I I have you know. I have a feeling that history will show that A. I was like the blind man looking under the Lamppost. His keys and someone came along and said where did you lose your keys He said well somewhere else. But this is the only place right can see. I was reading Donald BACI quote. I recently At the beginning of his book about the I which is just a fascinating area and I guess he spent a lot of his career and he did work in in the wool on radar and he was talking about the Radio Club. Which is these early Cybernet assist and the potential of the analog or digital computer to be what represented the brain and his perspective was he. He was sure it wasn't a digital computer and he wasn't sure it was an analog computer either and he thought it was kind of somewhere in between but it feels like that in between is what you're saying is that was the difficult bit to look and perhaps a police were able to look now. That's right I you know. It's I think it's being driven. This is true all science that what you cannot understand is is really determined by the tools that you have for making measurements for doing simulations in it's really only this modern era that has given us enough tools both to make progress with understanding how the brain works and also with a because of the fact that we have a tremendous amount of power now but just to go back to that early era. I think you know I once asked L. Annual you know who is at Carnegie Mellon and it was a time when Geoff Hinton was an assistant professor and I was at Johns Hopkins and I you know he was at the first fifty six meeting at Dartmouth or a I was born and I I said well. Why was it that you didn't look at the brain and for for inspiration and he said well we did. But there wasn't very much known about the at the time to help us out so we just had make doing our own and he's right. That was a era. You know the the fifties was kind of the the beginning of what we now understand about the signals in the brain. Actually potential synoptic potentials. So you know in a sense. What what he was saying was that we basically use the tools we have available the time which was basically computers but what they were good at. What were they good at? They were good at logic at rules. A binary programming. So that you know that was In a sense they were forced to do that. That's a really. WanNa come back to nine hundred seventy nine in a moment but this is an interesting context to that because of course. Vena initially was someone who spread across. Both these areas of Norbert Vena who was at mit founded cybernetics spread across both these areas of the analog and digital he did his PhD thesis on Russell and Whitehead's book but one thing I was reading about recently is there was a big falling out between Vina. I'm McCulloch Pitts. And it's sort of interesting. That Vena wasn't there at the I. E. T. in fifty six and I sometimes wonder was that more about personalities and wanting this sort of old guard to stay away because you always feel veto with someone who who bridge these worlds it. You know that's the fascinating story. I actually wrote a review of a book about Warren McCulloch came up. They were friends. They actually had had been friends yet. It has something to do with their wife's. Yeah I think the lifestyle McCullough was not line with its a side story but but I guess the point you're making which I think is an I'd like us to take us back to seventy nine and the meeting with Jeff is and I think that that's true. Despite the story between humans the real factor that drove things then was the sudden available at a t of increasing cheap digital computer. And no longer the need to do this work that Rosenblatt and McCain and others had done having to wire together a bunch of analog circuits. That you couldn't reprogram to build system. Yeah I think that was a dead. End It for the very reason you gave. Which is that you know you. It's a special purpose device. That isn't good for anything else. And and really if you're trying to explore you need the flexibility of being able to try many ideas and that's in that really is a digital simulation allows you to
Dating at a Distance
"Should PEOPLE STILL DATE? Everything is aligned. Date no no no no blind date center definitely not but even before this tender. I'm kidding that was mean on the late show with Stephen Colbert. A month ago before most of the country was staying at home at the time. A conversation about dating during a pandemic may have felt like late night comedy fair. You know Stephen. Everything in life is is a risk reward. Proposition is riskier to do things versus before. Perhaps being in close contact with somebody especially somebody. You don't know Is Is. It's a different time right now. But as their new reality has evolved so as the act of finding love people are using dating apps more both tender and bumble have reported an increase in daily messages and user engagement. Other APPS had a video chat feature and some people are reaching out in ways I would have never imagined a look out my windows. Bill dancing traps to take and needed to say here down. She waved back. That's the start of the story you may have heard before. It's from a video on Tick Tock by Jeremy Cohen a photographer from Brooklyn New York. After Jeremy Waves to the dancing girl he flies his drone over with his cell phone number. She picked up my job and I guess it works. 'cause I our lady's Jeremy's video went viral if we're still allowed to say that it has over thirty million views on talk now. I'm not at all surprised. It's the meet cute of our time if our time is defined by isolation and physical distancing Jeremy and Tori Cigna Rela. She's the girl on the roof have gone on a few dates after that. Here's Jeremy and Tori. The first date was we had dinner. There was another restaurant she is on her roof and I was on my balcony. It was so funny because we'd be talking to other on facetime and then sometimes I like look over like I'd see him there and then we'd look at each other. It was like such a weird scenario lovely on another date. Jeremy win inside a huge plastic bubble so he could take a walk. I just couldn't stop laughing. I like hit the ground. Basically I was not expecting to see him in a bubble. It's a lot more effort to go through than your average date and it's hard to express the usual social and physical cues when you're six feet apart but there things about this new normal that for Jeremy Cohen. Surprisingly work well. It's really nice to get to know her. Just not have any of this pressure at the end of the day like okay. Am I going home or am I gonNA invite her back to my place? This awkward moment of okay. What what is the other person thinking? I don't WANNA be too forward but I also don't want to be a scaredy cat. Jeremy isn't immune to the loneliness of social distancing of not actually being physically around someone even though he's found this new connection I am in my apartment either remained but he's with his family in Minnesota. So I'm alone in this two bedroom apartment for about a month. Now it makes me realize how much the small things in life such as a hub. Like hug skill great. I've actually putting myself a couple of times. It doesn't feel the same because it isn't the same. There's a lot of research that shows. That physical touch is important for health and wellbeing. One behavior that we have focused on in some of our research is interpersonal touch or affectionate touch. We've shown that touch has powerful effects on our physical. Health are mental health. Our relationship health. That's Professor Brooke Fini. She's a social psychologist at Carnegie Mellon University. She studies how relationships impact our health throughout our entire lives. It increases feelings of security so it just makes people feel more secure. It increases people's willingness to embrace life opportunities affectionate touch has been associated with lower daily stress lower reactivity to stress A lower likelihood of even perceiving something as stressful in the first place for Professor Feeney affectionate touch has benefits even above and beyond sexual intimacy. Which is something else were missing in? A time of isolation can engage in sexual intimacy for a variety of reasons that have to do with reproduction and drives and less to do with communicating care and acceptance and love and value. And so on. They're both important forms of touch and Communicate very important information to significant others But we think they are very different types of processes hearing about all the benefits of touch at a time when a lot of people are deprived of. It isn't exactly comforting. So what happens when we do lose it in our everyday lives? Here's Professor Brittany Kubiak. Who Studies affectionate touch in romantic relationships? Children form attachments with their caregivers in a lot of ways through touch and in adulthood we think that some of the same processes happens you form an attachment to your romantic partner just like you form attachment to your parents. Although the relationship is obviously different long-term not having the ability to touch. I think there's the possibility that you may not be able to form as secure attachment to that person but Professor Jacoby Act doesn't want to overstate the benefits of touch. Either it is still possible to have meaningful connections with each other without it. We know that people maintain very satisfying long distance relationships. Even when there's not a pandemic going on people do things reminiscing about times that they did spend together or planning times that they will spend together and so I think we can find ways at least if this is going to be a somewhat short term separation to make sure that we're maintaining high quality relationships even through physical distance for Professor Feeney. There is a positive outcome at least in terms of human connection about the fact that this is all happening to us together our rates of loneliness and social isolation even before they pandemic had been increasing and people have just been feeling more relational disconnected across the board. One positive thing that I think has come out of this. Pandemic is that people first of all are all in this together. You know so. We're now all part of a big group of people who all this happening to them. When people are facing adversity together they usually reach out to each other more and try to connect stance. Oh I do see one positive side effective at this is that there are these more creative ways that people are trying to connect and help each other out and so on like Jeremy and Tori and whether or not they do end up together doesn't even matter to them anymore. We're absolutely going to meet up. Probably something a little bit more low key like drinks but definitely still could never forget it honestly no matter what happens between us like we're going to remain friends like there's nothing like this that doesn't bond to people and said it's just like look if he's not like in my wedding he'll be at my wedding like that's. I certain that's a powerful connection. Now there was a study from Harvard. That came out this week. Saying we might have to prolong intermittent social distancing measures. Up until two thousand twenty. Two professor. Feeney doesn't know what that means for physical and mental health. No one does she and a team of researchers at Carnegie Mellon are just about to begin a study on this if we don't find these other creative ways to connect It remains to be seen. How well We can continue to abuse remote connections as a proxy for the more physical connections but I think the the core issue that underlies it all is. What touched communicates and so I think what we have to do. During the pandemic is just find other ways to communicate to our loved ones that were available to them if they need us even though we can't be physically proximal to them right now and might be more difficult to communicate that remotely but I think we can do it. Human beings are social creatures by nature. We crave connection. We're not meant to be isolated. These days. Some people might not have the security that comes from physical touch. But that's not all our relationship is built on find those other connections and lean on them. I think you'll be surprised by how strongly though resonate in your life.
Neural Architecture Search and Googles New AutoML Zero with Quoc Le
"Welcome to the PODCAST. Hi Everyone. It's great to have you on the show I've followed research for Your work for quite some time and I'm looking forward to digging into some of the new things that you're working on but before we do that I'd love to have you share a little bit about your background and how you got started working in machine learning okay so I was born in Vietnam. I did my Undergrad in Australia. And in my second year. My undergrad I started some project doing machine London with Alex. Mola a back in Australia and back. Then I was played with. Kodo methods Then I Did my PhD AT STANFORD. A on a lot of deep learning back in the day when deployed in whispers or very cool. And that's the route two thousand seven and around two thousand eleven I did a summer internship at Google and that was when Google Brin project was founded so when I was there that was a long and Jackie Naan Greco data was there and I. It was the sun so we started out small. That sounds cool. Yeah and then I did some of the Scaling Up Neuro networks with Google Britain folks and then You know at the end Up to two years did some work on machine translation with the media and Oreo VR. He's now did mine. Owner of Ilya is now at opening I and we develop sold end to end. Solution methods and Around two thousand sixteen. I started looking into more like You Know Auto. Mau Architecture search and more recently are looking to Malacca together with Otto may also look into Sent me supervised learning and it's awesome awesome now. You mentioned early on doing work with Alex. Mullah was he was this before he was at Carnegie Mellon was visiting in Australia. He was a professor in Australia. Yeah I I went to a university. In a small air. In the capital city Austrailia go Kendra. He was yeah camera and he was Professor Edward Research. So I thought I had. I have along Very interested in AI and machine learning and took me for that. I took a class data mining and so on and talk a little bit boring but the ability to actually learn. It's actually a super fascinating so I contacted him and he was moonlight co methods machine learning and we worked together for maybe a few years before he went to he went to America then. Cmu and Amazon. Okay okay so a lot of your. Recent work has been focused on this idea of You know automating machine learning and neural architecture surge to allow machines to find the best deep learning architectures in like. It's a little bit about how you arrived at working in that area. What some of the motivations were for getting started digging into that problem so I've been Along interested in this idea of self improvement machine should be self improving itself a machine learning and even and when I started doing co methods with Alex. I always ask him. You know how the Dakota bandwith and so on how some of the HYPOC Ramat does include methods decided and apparently they decided by using things like Cross validation on then where I work on. Koroma two narrow networks. My hope is to make the hype. Affirmative go away. But that's how is the opposite so if you look at the a Kabul Lucien neural networks at has a lot of hype privatised right like how many how many layers you want it to be and how many channels you wanted to be. And what are the some of the high assize apprentice since on a Coulda with all the training parameters? Yeah all learning. Dry and as researcher develop more and more techniques FAW EURONET. There's more decisions that you have to make. That feel like. This is like a problem that can be helped by a little bit of automation so So I I observe a lot of my colleagues who will when designing networks and I asked him about the principles of design. Your neural networks. And you started are having some really solid principles like Skip CONNECTION SO. The gradient can flow through the network concern. But as you tune the network Karen Hata do no longer have the principal is around. You know trial and error right you you try this a little bit and simply with better so you try that more so. I think that that is something that may be ready for automation so even during my Grad School. I already talked about trying this but I thought you know. Maybe we didn't have enough compute because training net already takes took me days so when I saw that new control. Units are are in thirty minutes. Something like that on on safer I thought. Oh maybe this is the right time to try this. So that's when I started doing this. Newer architecture search in two thousand sixteen. It's interesting that you know. Even with all of the compute resources of Google. You had to wait until the time was compressed. Enough in order to be able to tackle the problem. Yeah to get really good results. You want the networks will be really big and that will take a long time to train. Yeah and it's it's It's funny coming from me that we have so much resources that will go train in EURONET still taking a long time And so maybe talk about the the first steps in In that area. Did you jump right into neural architecture? Search or was that the you know a a an end stage or end result of this work where I I on some of the related ideas on and off since two thousand twelve like playing around with how to do. Better hyper profitable tuning and none of that. It's really published. Because I didn't have good results have pugh and so on so so I tried it on and off over the time you know every year I would set out some time to try this idea for a few months and you know and it didn't work very well because like a procurement song and then Two Thousand Sixteen. I met Barrett's off would as my colleague now at Google and he's very talented. So we say oh. Let's let's try at the idea of Jews in like a reinforcement learning to generate and network like a little layer in an network for for a ceasefire model. Seafoam motto. Is already at the time you could say that you know enough of you depends on how where you want to be but you from thirty minutes to a few hours and the seems like about the right amount of time to get this going and my prediction is that you have to train. Maybe either between from one thousand to ten thousand bottles and I did a backup our calculation and thought. Oh this might be the right time to do it but you know I have tried this some of these related ideas in much before
How to track coronavirus movement without violating your privacy
"With the corona virus pandemic exploding around the world. Some countries have taken too aggressive. Contact TRACING. Identify how it's spreading. There may be a way to do this without invading your privacy. I'm Roger Chang and this is your daily charge. What Me Stephen. Shanklin were the biggest brains at CNN. And someone who can understand when folks from MIT get involved with this project to get everyone in the same footing talk a bit about contact tracing. What is it? This is a very old technique for dealing with epidemics or pandemics where you basically A healthcare professional interview. Somebody who has whatever disease finds out who they've been in contact with that. Lets them trace? What the origin of diseases and also potentially figure out where it's going next What which people that person has been in contact with? Who might have got afterwards? So it's a very old technique for tracking disease figuring out who needs to be quarantined or sequestered needs to shelter in place or be treated and the problem is it doesn't scale. It's something you need a medical professional to do. And if you're looking at something like cove in nineteen the disease caused by corona virus. It's really hard to talk to. Every person in the world I mean this is affected hundreds of thousands of people. So it's not very easy to do. Large-scale contact tracing definitely. It's actually hit over a million at this point so Talk about this product. This basically like everything else in the world recruiting for it and so talk a little bit about this Private automated contact tracing or packed yet packed. So this is a project done by a team of people in. It is spearheading the effort but involves a lot of other universities as well including Brown Boston University Carnegie Mellon and some other folks as well researchers at a lot of different institutions. The way it works is it uses. Your Phone's Bluetooth connection to broadcast digital. Id Number from your phone into listen to digital. Id numbers being broadcast from other people's phones. There's no handshaking or or any kind of actual acknowledgment but each phone keeps a record of the digital. Id's it's come in contact with and what happens next is if you are if you test positive. You can voluntarily load the list of ideas you've come in contact with and if you haven't tested positive you can go to a central server and download the list of ideas and see if there's a match with your phone and that lets you find out if you potentially have been exposed. Gotcha and so this is. This is actually really interesting because the use of this digital ID. That's sort of the key here for keeping private right because with contact. Tracing I mean requires a lot of information. You know. You're you're you're giving all your details as well as all folks you've been talking to this sort of solves that problem of kind of keeping things someone anonymous right. Yeah exactly so there. There are two interesting comparisons. Here's the first is to traditional contact. Tracing a lot of people when they enter the healthcare system they might be okay participating in that kind of assistant. But if you're not you might not want to actually share all that detail But the other comparison is other ways of doing this large scale contact tracing with an APP so one obvious way might think about doing it is sharing a GPS log a of recording of everywhere. You've been your phone can very easily keep now. That's relatively easily done but you might not want to do that. Might not want to share that with the authorities are all the world so what the interesting thing is about. Packed is it. Lets you find out who you've been? Excuse me it lets you find out that you have been in contact with somebody but it doesn't tell you where it happened or who it was with so this let's this contact tracing happen in the large scale way without actually sharing personal details. It stand. Tell me some of the some of the folks behind this because there are some pretty big names behind his project right. Yeah there are some very interesting names To the people are Ron revest and audience. Jamir Those are to the people involved in the RSA Encryption Algorithm It's a very important invention that let's your computer. Set up a seeker connection On network so they are very big names. There's a big gap between having a lot of research credibility though actually deploying the system in the real world writing an APP writing out that works reliably and running the system. That does the records the ideas of people who have a negative. Excuse me running the system that records the ideas of the people who've tested positive. Checking it for security vulnerabilities. There's a lot of real world implementation details so these guys are really big names in the encryption world and they have a lot of credibility when it comes to things like security and privacy but they're still a long way to go between the initial idea and actually making something that works in the real world. Yeah I mean this is a problem that we're dealing with right now. Like what is for the time life or something like this because it is right now. It's a project. So it can get ruled out like are we actually can be used this in time to halt a spur of covenant or is this really for the next big pandemic. It's not clear at this stage. They don't have a release deadline they do have prototypes working on both android and us. They've actually had some trouble getting those two Getting smartphones from those two worlds to talk to each other but they do have prototypes working so they. They're not starting at zero right now but they don't have an APP released. They don't have a schedule. We've seen the new rates of couvert. Nineteen infections decreasing. So there's some evidence that the curve flattening is working. I still seems like there's going to be quite a while before. Were out of shelter at whom rules so there could be some window here for still for people to Get some use out of this APP and they're also could be. There's a lot of fear that there's going to be another flare up later after. Shelbert whom rules are eased. There's still a lot of risks that this could blow up again so potentially could be useful in detecting another outbreak. One of the really important things here to note though when it comes to the this time issue is how hard it's going to be to get this APP installed everywhere so it's going to be. We've seen cases where authorities recommend wearing masks or staying at home and then people voluntarily comply with that We haven't seen anybody recommend running a particular APP. Something like that could help encourage adoption or some direct support from apple in Google promoting this on their APP stores of their Kobe. Nineteen information pages. But without something like that. It's GonNa be really hard to get this APP into a lot of onto a lot of phones fast right. I think. That's that's my. That was my next question is really the scale ability issue like this only works if everyone download this. App EMBRACES THIS IDEA. But that's not the case right. Well you can look at it two different ways. It only works in a large scale way if a large number of people download the APP but even on a small scale could potentially inform some number of people that they are at some risk of infection. So if some small number people are alerted and get follow their symptoms more closely or potentially get treatment. It could help with that small number of people but for this to be really effective at the large scale it yeah has to be deployed widely. That's much more challenging. And is this the only game in town? I've seen pitches for a number of other APPs that that promise contact tracing that prompts the track things like what's the difference between those bitches and what you're talking about today. Yeah there are several efforts actually some of them have already come together. This one At Mit joined with some folks at Boston University. Who has very similar idea there other efforts? There's one in Europe. There's something called Kobe Watch. There are different efforts that use this basic approach than there also are other efforts that use other approaches and then there are other apps that do things like help you identify your check your symptoms and other apps that I've been used to see if people are generally obeying the shelter in place requirements or advice. Lots of different APPS. It's very confusing right now. I think one of the big challenges we're going to have is for people who want to help themselves or help the overall fight. It's going to be a kind of a mess to figure out what. Apps they're supposed to install what APPS are useful. What apps are potential privacy invasion and? I'm sure there can be a lot of cases. What APPS ARE SCAMS OR GARBAGE? Yeah that stuff is something that's GonNa Watch out for.
Coronavirus Detected By Voice? Carnegie Mellon Researchers Develop App To ‘Listen’ For Signs Of COVID-19
"News researchers at Carnegie Mellon University creating an app they see might be able to tell if you have covert nineteen the app hasn't been approved by the FDA or CDC and is still in its early stages it listens to you cough and then asked you to resign and number of letters and sounds before letting you know how likely it is that you may have the disease as of now researchers say the app is an experimental and
Google asks users about symptoms for Carnegie Mellon coronavirus forecasting effort
"Google is working with Carnegie Mellon researchers aiming to forecast the spread of coronavirus infections Google says over the last three days it has surveyed some users about their health at the university's request researchers are receiving the aggregated and anonymous
The Third Wave of Robotic Learning with Ken Goldberg
"All right everyone. I am on the line with Ken. Goldberg can is a professor of engineering at UC Berkeley. Ken Welcome to the TWAT podcast. Thank you pleasure to be here. It is great to finally get you on this show. We've been talking about this for a bit. You know I meant to ask you before we started last time you were. You mentioned you. Were working on a book. Maybe we'll get remembering that right. Well I think I'M I. I've been thinking about that for a while but I'm also thinking about an more right now an article. Okay okay. Well we'll We'll get to the article. I think I I came across you and some of your work in the context of decks net. I saw that at a Siemens Innovation Fair last year. I think we exchanged tweets and stuff like that. But you know I would really love for you to introduce yourself to the audience and share a little bit about your background and how you came into working in robotics and okay great. I well first since you mentioned twitter I should mention my twitter handle which is at Ken. Underscore Goldberg. And I've been trained very well. My daughter to post there at least one today so I've got the actually. I found it very interesting channel so so I am posting technical things as well as updates about things that are that are finding out which is the learning about which I find very useful. So my background is that I was. I went to University of Pennsylvania and then went to Carnegie Mellon for I was at USC for four years and then to Berkeley where I've been for now twenty five years for here. I RUN A lab. The we we call it the auto lab for Automation Science and Engineering and we have approximately thirty students doing research in there. And we're doing work. There's there's there's post graduate students and a good number of undergrads and we're also associated with other labs like the Berkeley Research Lab and the rise lab and citrus and other programs at Berkeley our particular labs interested in in in doing research on on robotics basically on algorithm ick approaches to robotics and specifically in last year's been focusing on learning methods for for imitation learning deep learning and reinforcement learning for control of robots in applications from grasping as you mentioned which is a primary want working on for for thirty five years to surgery surgical assistance Hugh assisting human surgeons for for robotics and home robots to especially for seniors and in who are who are who prefer to live at home and the last year is very new and we can talk about later is is agriculture and we have a new approach to poly culture farming that were exploring using deep learning so one thing that I thought was really interesting in looking at your bio is in spite of the fact that you are a highly accomplished robot assists you start your your body starts with Ken. Goldberg is an artist so art clearly must be very important to. You actually saw some sketches behind. You am curious. I'm curious about Ken as an artist. And you know how if all ties into your work. It's not the usual fare of this podcast but then I saw somewhere else. You are filmmaker as well Is that your art? Tell US okay. Well actually I wanted to be an artist when I was a kid and I I basically my mother said listen. You can be an artist after you become an engineer. So She she. She was very wise and I think it was. It was it was a good choice for me because actually love both art. Something that I take very seriously. I think it's often underrated by many people especially Engineers who think of it as as lightweight. It's actually just opposite trying to produce something that's meaningful in the art. World is extremely difficult and demanding. So I've spent a lot of time studying I have made a series of installations and projects. That almost always involve technology in some way. But they're also commenting on the role of technology in society. So probably best known pieces of project is a project called Tele Garden that my students and I set up in the very early very early years of the Internet. So it's nineteen ninety five that we we connected a industrial robot arm to the web interface at the time which was mosaic Browser and we built an interface. That would allow you from your screen from anywhere from your laptop There were no cellphones at the time. But you could. You could log in this thing I think. Yeah it was very fun project. We thought well. It's kind of curious. who would use it if anyone and we got thousands of people coming in and and moving the robot but the part of what was made. An artwork was the context because it was sitting inside a garden. A real physical gardens. We could plant in water seeds remotely and then we got tens of thousands and we estimate that over the time that product was was that robot was available online which is approximately nine years. It was visited over. Hundred thousand people participated in the in the project. That's awesome that's awesome again kind of the technology and are coming together rate. So that was the thing Sam because one of the ideas were said I. I don't think I would have pursued that if I just stuck with my research plans at the time but because this came out and offered a way to reach a at the time when I saw as potentially very broad audience I started putting effort into this then there was a fantastic team of students who worked on it. And then we are thrilled with the the idea that you could take a robot and you could put it into the hands essentially of potentially millions of people and then there were. There was a proof of concept the interface questions there it turned out that there were lots of interesting theoretical questions that came out of that so after that project we did a series of subsequent projects and then had an NSF grant to develop versions of this. We have a patent related to the south. Yeah it really grew into a whole new direction of research that that really started with our awesome awesome into tell us a little bit about your research interests nowadays more broadly. So we're still doing art and I can come back to that. There's a new contract. But the the the lab right now is been been very very focused on robot learning and especially as as I know your. Your listeners are very aware there's been huge revolution in the past decade. And so we've been. We're interested in this before the the advanced in deep learning started but now it really has become a huge focus for us so in particular. We have this been working robot grasping for many years and then went deep. Learning came out. We saw an opportunity to apply it. I can tell you that story if you if you like how we do it. Maybe start from the perspective of the grounding on the challenges associated with grasping like we see these pictures of whether they're rohbock robot hands or more industrial types of robots or prostheses. And you know a can grasp like we've seen we've all seen pictures of that but maybe it's harder than it looks or you know maybe the opportunities that have not figured out. Oh good okay so I can. I can answer that partly. I've realized only last few years that part of the reason I believe I went into this field was at myself as a kid was was incredibly clumsy. I still I still am. Anyone would throw me a ball I would drop it and so You know the last kid getting picked for any sports Games or anything like that and it was just Ed. I think that may unconsciously made me interested in in trying to figure this thing out like how. How do you grasp things and many years? Later when I was in Undergraduate I joined a laboratory of the University of Pennsylvania and they were studying various aspects of tactile sensing and I built a very simple hand with another student and we started really exploring this question. Of How do you grasp things? And it is fundamentally difficult for robots like to say that robots remain incredibly clumsy today. They're much better than they were but industrial arms. If you give them novel objects there will drop them with a fairly high frequency and this is a problem because we really want is want You WANNA be able to pick up anything that's put in front of you and the application the big application that's growing enormous right now is e commerce so you wanna be able to take objects every orders different so you wanna take things from bins and pack them. Lift them out of the band. Grasp and put them into boxes or bags for shipment and that turns out to be a bottleneck right now for robotics
So We’re Working From Home. Can the Internet Handle It?
"And with millions of people now working now taking online classes and sheltering in place at home the next few weeks could put quite a strain on internet service the big question it can the infrastructure handle it for a closer look Stan Bunger and sizzling Taylor spoke with John piazza professor of Kerr at Carnegie Mellon University and former chief technologist with the Federal Communications Commission can our infrastructure handle it that's the big question well the simple answer is we don't really know nothing like this has happened before luckily I I think most of the usage will occur during the day when it's not peak hour there is excess capacity so if we're lucky we won't exceed that capacity but we might especially in places like the bay area will just have to wait and find out until now the assumption had been that you know drive time as it were for the internet was the evening when everybody was busy streaming Netflix and Amazon movies is that still the case well that was the case two weeks and that that's why we have excess capacity during the day because they build to meet the the peak our needs we just don't know how much the daytime usage is going to go up if there are problems what would the first signs be or will it just stop no I won't stop you'll just see poor performance yeah for those of us and now I have to teach my classes hotline for those of us who are home working or or doing other things we may find that video for example doesn't work and if you're on a cable system particularly in the upstream is a problem if you are video conferencing the video you're showing is more problematic than the video you're seeing and we might have to go to audio of that half yeah and to be perfectly Frank just to share a story here we may have run into that very problem here KCBS yesterday morning when I tried to set up with a high quality audio codec from my house and it may well have been a limited upstream capacity from my house that brought things to a halt it's very possible so the question of which in the walls of somebody's house we're all on the same wifi network in any family structure other issues there too there could be we particularly actually the your in say an apartment building with ten walls and and and are densely packed you may actually find that the neighboring apartments start interfering with each other in terms of wifi that's also possible do you know if providers are doing anything to ease the situation I I mean they're they're they're changing some some of their their offerings to customers I'm actually trying my research team tried to reach out to some of the I. S. P.'s to to in the hope that they will share data with us particularly in the places that are getting hit first like Seattle and and now the bay area so we can learn and help the rest of the country prepare but if any of them are listening I hope they'll call me the cooperation that was John P. hopp professor at Carnegie Mellon University and former chief technologist with the Federal Communications
Visualizing Fairness in Machine Learning with Yongsu Ahn and Alex Cabrera
"So let's get started with the with the topic of today so today we talk about a really really relevant topic can needs It's particularly hot right now. We're GONNA talk about bias in fairness in machine learning. And if you know know what this is we're going to describe and explain what this is about in a moment and more specifically what is the role that can play in this specific domain to say mitigate problems that can arise in terms of bias and furnace in machine learning so to talk about this topic. We have not one but two guests. We have Alex Cabrera who is a PhD student from Carnegie Mellon University. I Alex is again. Thanks so much for having and then we have young. Soo on who is also Ph student at the University of Pittsburgh I- youngster. Welcome to the show. Hello Nice to talk to you so Alexander. Who can you briefly introduce yourself? Tell us a little bit about what is your background. What is your main research topic? And just give a brief introduction. Yeah so I'M ALEX. I'm a PhD. Student of the Human Computer interactions to at Carnegie Mellon so generally idea research into creating interactive systems and visualization. Systems that help people both develop better machine learning models so even more accurate more equitable and understanding these models so understanding potential issues. Or How? They work okay. Young Soo my name is sue on in I'm a dirtier peachy students at University of Pittsburgh. A my research interest lies at the intersection of visualization and fair. And explain away. I enter to machine learning so my primary research question is to build assistant to help users with making the machine learning results more fair and explainable in helped him to interact with machine so that their opinions and apex can be incorporated into the system. Okay thanks so much so I was thinking. Maybe we should start with defining a little bit this terminology to the extent that he's possible but maybe they're probably many of our listeners who've never heard of that and of fairness and bias and this is a very overloaded terminology here so I'm wondering if we can start by defining a little bit. What what we mean by fairness and maybe even bias in emission learning and also what? What kind of province exists there yet? So I'll probably I can start by talking about a little bit of background on why the problem This fairness problem has been actively discussed in especially missionaries research. So as a May have seen did. Data driven decision is kind of increasingly used in important decisions so especially Such as a job recording of colleagues dimension were predicted policy. Those kind of important decision which have kind of huge impact on Individuals muster learning as more and more used induced kind of important decisions then Some of cases have been reported that these machine turned out to be biased towards certain groups or certain individuals so here the what I mean by bias is certain. Decisions are kind of burrow favored to certain groups or individuals. Such as man over woman or a white people over african-american people. This is because on the machine. Learning model is trained from Historic Co. Data set and this historical data said could possibly include Inherited bias then. The model is kind trained by those data sets and then have kind of inherited vice. The problem of machine learning here is that whatever trained model can kind of systematically discriminate against certain individuals groups especially in Western Assistant Because many decision makers may use to system in their decision making then kind of making these mistresses. More Fair is kind of important problem so basically the type of fairness you talk about is mostly related to not being discriminatory or not using features. That have nothing to do with the essential decision. You're making more superficial like Maybe the race or gender or other features of a person right. So it's about combating discrimination. Yeah I think that's the main idea. It's actually you get to you a more complicated because even if you don't include some of these protected features so if you say you're trying to give someone alone you don't really want to decide that based off of their gender their race Those are actually. You can be almost perfectly predicted by the other features so you can actually reconstruct that so actually a lot of machine. Learning people suggest you actually add those features in because they're going to be used anyway and then you can apply some resolutions afterwards to try to address the problem. So it's very much embedded in the data that you're using to train the model this historical data that you've collected so it's not just as easy as leaving out that column with Race Agenda and not saying you talk to the research that's happening now. It's a little bit more complicated. Just the complex relationships between the variables ends up that you can actually recreate the biopsies. Even having no idea algorithm not being aware of these protected attributes. Okay but just on the Senate so the evaluation you do and fair. Evaluation is one that only takes the features into account that you're supposed to take into account so usually the way we did try to define fairness or quantify is an output. So if you're trying to give loans or a very popular example is trying to decide algorithms to decide how risk how likely someone is to recommit a crime if they're like. Oh so whether or not you should give someone bail we usually it doesn't we don't really look at what features are used that we look at. What the output is and so if for example the RECIDIVISM prediction case for African American males? You're more likely to be given a higher risk or even though you're just as likely to recommit a crime that is discrimination. That is the bias that we're trying to discover and trying to combat right So we really like black box models. It's really hard to know. What parts of the data are being used to make the decision? But we really care about whether these decisions were making. The outputs are making that really society impactful whether those are equitable and fair. Okay Yeah I'm wondering if we can can you? Maybe describe one or two specific examples. Where these kind of problems can arise. I think what is interesting? Is that right now? I mean we live in a society where where these these systems stems are already making decisions or some of some decisions for us right or providing indications for for experts that have to make decisions based on on what the AI system suggests recommends right. So I think I'm wondering if we in order to make a little bit more concrete if you can cite one one or two examples where where these these. This kind of problem can rice yes. Sadly there are quite quite a few examples. So one of the biggest one of the first investigations elected to it was in facial recognition systems so there are systems by like Had some face plus plus IBM and Microsoft that they audit and it tries to tell given a picture of someone's face whether they're male or female and when they started looking into it they found that when you start seeing how well they perform for say white men versus darker skinned women. There was almost ninety nine percent accuracy for the white males and close to seventy percent accuracy. For the darker skin females which is pretty big disparity. A lot of that is due. Hey if you look up. General data sets of faces a lot of the faces. That come up are white males that data that you're learning on is not
"carnegie mellon" Discussed on The Knowledge Project with Shane Parrish
"So if all do all teams have the same information do they all use it the same way like where's the edge in terms of analytics, and who does analytics really well versus like who's still learning you know, again, I, I would never apply about somebody else's organization. I think we're very fortunate in Pittsburgh. You know, we're sitting next to Carnegie Mellon University. Which is one of the top in the world for analytics machine learning AI. And so the person who runs our analytics division is from Carnegie Mellon, and, you know, we have tremendous resources there and is that the future of sports and are those algorithms or sort of insights proprietary now more than the east to be. Yeah. I think so. And I think that. It really is a blend of art and science, though, because. You know, you still have to be able to think about how again, human beings are gonna function and leverage situations. And you're gonna have to, you know, I think the highest and best outcomes are where you use the information and the technology. But you're able to interpret it in apply instead of just saying to inform your decision. That's exactly right. And I think certainly you look at baseball it has become an arms race. And for the teams number of years ago that rejected it and say, we're not doing that. Well, you know it's tough to be planning October. When, when you don't have the right amount of information, and certainly baseball is easier to apply. Analytics because you have a lot of one on one. Exactly. So it's it's easier to apply. But football's come a long way. Do you think we'll ever see like a? AI version of an offensive coordinator or defensive coordinator. Yeah. I mean just from calling plays not from coaching perspective, but it's like this play this. We know what's going on on the field. We have real time data, or that would not surprise me. I mean, or or at least, you know, like Jarvis in iron man, right? Like, here's what you should do that, that to me is not far fetched at all. How far away, do you think we would be from that? That's an interesting question, because I always find that technology you lose usually happen slower than you think. And then suddenly. Because if, if, if somebody used it, and it worked, it would be like nobody's wearing headsets and able to communicate upstairs, and then all of a sudden one person does, and that's an unfair advantage. So I would be completely guessing. But I think over the next five to ten years, you're gonna see remarkable advancements, you know, across the board, and in artificial intelligence. So are we had to play where we're using intellects to adjust players practice routines in the sense of like this player worked really hard? We need to give them more of arrests this player can work harder. They can go like more wraps, like, how are we using that to inform the individual players outside of games? I'm not sure we're quite there, yet there are tests where, you know, the guys you spit in a Cup, essentially, and it says, look, you're, you're low on x y and z should hydrate or you should so. I know there are clubs using some things like that. I'm not aware yet of using that level of granularity at practice by do you think that's what we're going to or will always just be this element of, I think, in the foreseeable future. It will be a tool in the toolbox, but not you know the tool. Yeah, I don't look like anything else. But sports sports is a great imitator. So if somebody, you know, employed those types of things in won the Super Bowl or won the World Series, I you can, you can bet that people are going to emulate it. What's the difference between without getting into specifics? But the teams that are perennially good in salary cap era, which is hard and teams that sort of, like build up. They have one or two years of success in the fall off the clip. Is that all salary cap management or is it? Man, I my own personal opinion is, it's a number of things again back to that word of culture, you know, do you have athletes that wanna play for you? Wanna play in that city?.
"carnegie mellon" Discussed on KTRH
"I met him at a UFO conference in Washington DC, just an amazing guy. He passed away recently at the age of eighty five years old his academic background included a bachelor science degree in industrial management from Carnegie Mellon University. A bachelor science from the US naval postgraduate school a doctor of science and area. Nodding astronautics from MIT. And he was a staunch believer that this planet was being visited by extra terrestrial head girl. Pleasure, my friend. How are you George? Nice to be with. It was morning. Likewise, likewise, you've been busy lately since we last talked stay. I stay pretty busy going on. Well, I had an opportunity to meet the Buzz Aldrin. I was king show several weeks ago. And what are the the light full guy? He's he's tough. Now is he probably was then at her. Well. All right. He does not believe in extraterrestrials. Did you ever talk to them about that? Briefed. I've talked all developed about what they thought face. But. No, not obscene pretend experience for the Apollo program. Nothing in space was subsequently validated. Are understood. Good,.
"carnegie mellon" Discussed on Newsradio 1200 WOAI
"Let us know what that services, and we will get our content up there, and we will get you a cyber talk radio T shirt. So yeah, you've definitely came across project quest. They've got a long history here working with workforce development in the bear county area and have done an excellent job and have started some cyber programs. I'm glad to hear that they're taking things here of to the next level with an apprenticeship program and moving beyond just the facilitating of the education. Since we've joined with them. We've also gotten them to be a part of. Nice which is the National Institute. And it's cyber that helps us that is helping set up programs like this apprenticeship program across the nations, and that's one of the bigger pieces that they're working on. And we just have. Right now, we have two programs one shut up in Peoria, Illinois one which is starting here this month. And. San Antonio through Shanto college, and we have seven locations in the states on the east coast that are becoming a part of the process, it it takes a while to build the process if you want to know a little bit about what how we do. The software the secure software development. We we have our software programs come basically from Carnegie Mellon University. And we've been in an alliance with them for several years, and we have. Good association with the security engineering or software engineering institute at Carnegie Mellon, and they have a certain set of programs which they've done a lot of work on that started at masters, then it bachelor's and now headed apprenticeship program at associate's degree where they just describe what they think is the best curriculum for people so sack has pretty much a dentist and adopted that program, and we also are coordinating some additional training for sack professors that will be part of this program as well. You're listening to twelve hundred WAI this is cyber talk radio and we're discussing cybersecurity apprenticeship program. That began originally up in Peoria, Illinois, and now is become added. The second location here in San Antonio, Texas, working with San Antonio college is Texas's first cybersecurity apprenticeship program, so. And specifically in cybersecurity and secure software development. So the buddy is is you guys were going to build this program here over the last eighteen months in Texas, where did you find students for the program? Pretty easy. We were we actually were didn't really know where to go. What to do? But we shut up a couple of. Conferences with some professors at sac and. We went into shock and made a couple of did a couple of briefings to a couple of computer security, your computer programming organizations that were full of students and the first day that we talked nine of the guys swapped their major from whatever it was to secure scared program programming. Curriculum. And and we've kind of been fluctuating back and forth as as we get. This thing started we have we've had some stops goes, some delays and. Now, we have gone through the process, and there is a true process for selecting students. They I have to be enrolled at sack. That's the first thing. The second thing is they have to take a. Even though they may be in their first year deep into their first year what their role they have to take computer aptitude test and make sure that it is what it is. And then. Quest goes to up their process with them, which details checks and balances such as that. And then now we're in the process where we're going to be introducing the higher proposed apprentices to the employers. So the employers get a chance to interview them and get to pull them in and hire them if you will. And that that's basic. We've talked the question talks to a lot of veterans and openness, we've opened this, of course, to veterans into the people who don't often think about wanting to program or or maybe have a little bit of experience in programming or I've taught themselves to program little bit. And they don't know where to go or how to get there. And we're basically bringing those folks and explain to them difference between somebody who just learned to write Java. And now they're now they're a programmer and how the process works. And and I'm kind of stilted at times, I say that. In the normal world people will write for X period of time, and they think they're programs ready, and then they turn it over and try to run it under operational circumstances. And they discover that it has a lot of bugs, so they stopped that'll stop. They continue to run the program and fix the bugs as they go until they get to the point where it is operational. And if you follow the process that we use you get to skip that part because you've already worked all your bugs out when you when you call yourself through and you just go directly from producing the program to operate in the program and doing care and maintenance you stop to care, maintenance and updates on it. Yeah. And so for the selection into the program, if you're if we have students at stack that are out there listening right now, and they go, hey, I want to get into to this program where do they go to learn more about it? The easiest thing is for them to reach out project, and we have three ladies over there who are very heavily involved with this activity, and we that's that's the easiest thing to do. They can't speak with their professors. There's a couple of professors, but I don't I'm not I'm not sure that I really need a bargain to say user names, or whatever because I don't have that permission. But there's a couple of professors that are out at out there that are heavily involved in this program. Project quest great front door. Go talk to them. Even if you are not enrolled at Sackett, I would say go talk to project quest. They'll walk you through what you would need to do to get an enrolled at sack what you would need to do after you are enrolled to apply for this program. So they will be great front door to get. You have all that information that you need so from enrolling in the program. How how does this apprenticeship work with the schooling and working for the employer? How does that portion of the program kind of run along after a student is selected either end of the program by the school and then and then picked by an employer. We start off with students attending school. And then we'd go through the hiring process. We pick up the apprentices as as employers as and we you know, we go to the novel 'em processing statements that we would do. And and our way of doing thing our rules each company, separate and different. There is no there is no pure organization such as that goes. But. We work. We are set up to work are allows an apprentice a student to work for between nineteen twenty five hours a week and part of that time frame that they are going to. A working. It can be whenever it is that we need to be as long as the companies companies stood at worked it out there. So there's no provision for that. The student continues to go to school at the same time. There are programs. We our program was set up as as blocks whether you worked X number of hours. In the company, and then you move when the time was right? You move back to the block of education, but we're doing the I call it co mingling the to the two activities together at the same time and it works. We think it will work very well most everyone was in favor of that of that activity working that way. Do you give them a little bit of experience in and writing? And then bring them into the company now in addition to when they bring them into the company. We are certified and the SEI program obscure software development, and we have an agreement with the other companies that at part of the time when there I'm gonna say probably four hours a week part of the time when they're working at the companies will pull them into a either a once one position place are pulled him onto the web web systems, and we will run some training for them that comes out of SEI that particularly teaches habits. Teaches. Good. Good habits. Vice bad habits teaches how to. Get along get with the other people in become and work as a team. And then we do the strikes what I call the software security pieces that are produced by SEI because at the end of these programmes we have to search setup. We have a shirt that's called an associate cert- under IS square that she L S P, and it allows an individual who's basically brand new in the business to come up with a an associate she S P, which is a really high level cert-. The only reason it's associated as they don't have five years experience when they reach the five years experienced, a they go over to the straight over to the SEAL as an FBI also has a certification, and we train the people to work under the SEI curriculum. And then we go over and work hard to ensure they're capable of passing the SEI shirt. So when they graduate they ended up with two search nobody else has right now. People when you walk out of college. You certainly don't have and and then coming out of the military. You certainly don't have those to the main places where people come out for. For software programming. Yeah. And for those certifications for folks that are listening and wondering they're valuable for private sector. Employers are certainly valuable for if you were in the military, and then you wanted to continue working after you you transition from enlisted to a contractor. There's many contracts out there that will maybe either recommend or require those certifications for folks that as you're writing software for our government. It's being written in a safe and secure way. Right. And that's that's the most of us are government contractors at this point in time. But we actually looking for a couple of companies who are what we call shebeen companies. That are not in the government software business for so for a company to apply. So we've I'm sure everyone out there right now is going man. I wish I could hire software developers one. And then if I could get a software development accent knew how to write bug free secure software. I really want to hire them. So where do they go to to get involved in this apprenticeship program? Most likely the best places to come to me. And I I'm very easily reachable. But so people would is best that people would come to come to me. And we'll set them up. They can also go to quest and talk with quest quest as a direct line. And as is perfectly capable of fulfilling telling them to give them any information they need, and we will we are. We're very much looking for additional companies because the fact while we have a small cohort starting right now, we expect that in January we're going to have one that's twice as large, and we need places for the for the employees for the apprentices to be employed. So for the company, what are the asks of them, what requirements are they committing to through this process? So I mean, what I've I guess I've heard so far you're going to have a part time employee for a while while they're in school during this apprenticeship how long does that last? Are there any pay guide? Lines. How how does that process that work here? From a training perspective. Okay. There is a there are pay guidelines. First off we start right at twenty dollars an hour and over a two year period. Because that's what the apprentice program is we raise that program Russia shower up to about fifteen dollars a little over fifteen. And that's where that's where the all the apprentices when they quote graduate will will be. We. It's a two thousand our program, and that's shut up by department of labor because we are certified by the department of labor, and they. The program is is called a mixed program because the fact is education, and it's showery. So it's about half and half a number of hours in school number hours in working and from that perspective, the companies are required. If they hire someone they required to keep keep them and pay them and help them learn and grow and the benefit is, of course, is that you have somebody to work with for two years to figure out whether or not you think evade employees. You have a lot of people that. Can come across. And and be very productive after the first year as opposed to just being employed as an apprentice for two years, we've had about half of our apprentices that we that. We brought in who knew nothing or some who knew a lot about after the first semester in half to two semesters. They were being productive and time chargeable. Which is to be able to do, you know? What the other piece to the company is is that we do spend a lot of time teaching the SEI programs. And so there are stipends that to the company that helps cover the cost of the of the program. We by no means for making profit on those on those teaching that software. Yeah. But we we do charge a stipend and right now. Stipend for that. And some of the tests and some of the other activities this actually about seventy five hundred dollars, but it is considered to be a one time stipend, not every co heart not every piece, but one time when you come in because it helps cover the cost, and we're developing ways through grants, we hope to eradicate that entirely. So that's where we are. Right now, though. So you're listening to twelve hundred w e I this is cyber talk radio and buddy Smith, and I are talking about Texas. First cybersecurity apprenticeship program, we're going to be taken a break here at the bottom of the hour.
"carnegie mellon" Discussed on Recode Decode
"Yeah. If you look at the meteor, it's like between learning this and intelligent agents that right, whereas now as a people are going back in the hands to the electrical, computer engineering, this happens at Comey amount on its books in EC primarily helping us solve the sense our. Yeah, that's that's one end. Yeah, there's another end above machine learning. If you go something which can spot patterns or notice that your elbow is in is in an image and stuff like that. You still going to want to put it in its system which makes decisions. Right? And that really is the original dream of official intelligence of the dominant conference these things which can observe think about what they observed and then act. Right. And so making the action decision is really important, and it goes into ways. One of them is a little bit like warrior talking about with. Google. The early days of this is a human has to make a decision and as much too much information around for them to actually really be able to just look at all the source information themselves, how can you support that? And that goes from everyone doing stock trading to helping decide if a medical claim is legitimate through to people. One of our professors for instance, is interim instrumentation classrooms, so that teaches can notice if they've accidentally got some unconscious bias, which means that they're not attending to kinds of students and say full. So that's great. That is human assistance. In my opinion, many folks in the official industry intelligence industry. By the way, I'm a minority. I not speaking for the whole discipline focusing on that because it's so much more palatable less scary than the other thing at the top of the stack, which is autonomy. Right? And. Nope. Meaning to him. Ever ever write us robots, don't care about. That toasters toaster is one of the other. They don't. We don't. They don't care. You know, oddly ahead, it didn't in a good deal on must have been using. I thought he was quitting tells you that they're going to think of is like house cats. Like, why did they want to kill us? It's kind of useless to wanna get like it was really, I was like, oh, yeah, you're right. We're. So we're so obsessed with ourselves into science fiction feeders territory. The reason they don't care about us is then thinking about us, they are simply machines every row books, computer that you can see is just a set of. We only have just a few minutes just two more minutes. I just love to know why we're self-driving is because that's where community you were all involved. Lots of people left Carnegie Mellon. They set up different shops. Where do you see that right now self-driving Thomas vehicles. This is me speak yet myself. Yeah. So my personal understanding is that every year now in the major self-driving experiments, the metric of success to track. Is what fraction of the time do you need a safety driver? In other words, what fraction of the time does the human need to take control? And if we were shooting for the early two thousand twenty s for us to be at the point where you could launch autonomous driving, you'd need to see every year at the moment more than sixty percent reduction every year to get us down to ninety thousand nine point nine, nine, nine, nine, percents safety. I don't believe that things are progressing anyway near that fast, right?
"carnegie mellon" Discussed on Recode Decode
"So there are many creative things that can be done to reduce the cost of these kinds of of education, Georgia techs, mostly online system. As an example. That's I think that's a place where I can demand has to go now is to actually value engineer. It's expensive master's programs creatively so that people can afford domestically. Take me. Get you have all this demand who people who can pay in. That's where you have. So I'm gonna finish up talking about the concept of where the big trends in academic sir, going around computing. What do you see is important you obviously robotics and. Missing saying that Cardi Mellon was a big player in Uber's Uber, a lot of self driving cars, a lot of robotics, a lot of things like that that had sort of rocky ten years. Let's play mover for that, but talk a little bit of where it's going, Carnegie Mellon where overall, you see the most important areas of computing going. The huge change. Of course, we will seen in the last ten years is being machine learning and the the real push on these convoluted neural networks, ton outs to Bill to solve problems that we haven't been able to solve using most of his super AI in Silicon Valley. Now, whatever the new marketing turn on the radio. But if if we went on the radio, you'd have seen me, I rolling. They have all kinds of names for it. Go ahead, just marketing, I don't mind. So he's, he's the important thing that machine learning component fits in a slot of all the technologies. You need to build an a system. So one of the things we've really been pushing on both in our developments of education and in our recruiting faculty, the other slots adjacent to machine learning. Right? So important one, which I've usually withdraw before machine learning because machine lending depends on it is all the sense work sensing work this necessary to be able to understand the world, right? So that that's an astonishing area, but exactly. So if you're using robots to fight a fire, they absolutely need to understand what's really going on in the building. And so creating devices and robots, which can actually understand what's happening right now is I think if. If I only had one research topic that I could work on, I would regard that as much more important than improving the algorithms which is going to take that sensory data. It's going to be amazing when they figure it. I was talking to someone one of these future. I like to talk to a futurist and they were like at some point. And I think it's actually being created, say there's a nuclear spill or so the chemical spill or something like that that they would have sensors that were small as like grains of sand and they throw them on like from faraway, they'd spray them onto something in these sensors would pull in and the information of what spilled and what to do about it. And it was, I love the concept of it like, I'm sure it's not. Possible at this point, but that's the idea. It's like it's there. So sand is the way I looked at like they're so pervasive. They're almost like in the air without knowing their there. Yes, I do think we're moving in that direction. It actually totally makes your totally making this point. The. The idea of just trying to really focus on machine learning without being able to get hold of the killer one really important part of that turns out to be power and actually having it so that a sensor which is tiny is purchasing high-definition video. You can't have it sitting next to a big GPU of the now putting in. 'cause because that would be all the weight for a mobile platform and so forth. So for me, a lot of the stuff that we mathematicians like me doing in the middle of it movie stalled without that huge growth of book in the which he had our sensing and sensing everywhere going out to space going out inside of people. I mean, I remember that movie where they've traveled inside a human being. I'm like they're going to have sensors all over human beings at some point if we can deal with these things. So a very interesting aspect of that is, in my opinion, electrical and computer engineering departments, which. To some extent of had to watch computer science, getting older blurry that hopefully coming up. Oh, man. With friends, but still wrenches going what the hell if you used to be cool..
"carnegie mellon" Discussed on Recode Decode
"Taylor, right? I'm many people had direct or indirect friends who are running small retail business and very concerned about how they can make sure that that what they're doing is visit night. So a lot of people in the Amazon sneak up on them that was like they're behind your back. Well, thinking like, oh, they're our friend. I'm like, they're not your friend. Yes. And this is not your friend? My my understanding is that through many of the larger retail companies, this this is being a big looming thing. Different retailers have moved at different speeds with saying, look, we've got to seriously invest in a cloud online offering, and for many retailers, there has been this question. We'll maybe we shouldn't do that preps. Our whole strategy should be about walk-ins and people have any experience doing the physical shopping. Right. So this question of what to do as during a disruptive period and shopping has been sort of raging on people really interesting where we are now. Of course, there's I, I. Again, as I said, I used to say the calls coming from inside the house retailers. You just like you're going to die. You don't be careful. So you left, then you left Google to do go back academia, and why did you do that? Because a lot of people like to stay in the fray and obviously a lot of academics also spent a lot of time in the fray. It was. It was fascinating. What I noticed going on in the world of academia was it's really centrally important role for the economy in for the future. The one of the biggest trends going on among the tech companies in what's going to keep them alive is how much strong software and machine learning AI expertise they can get hold of. And that was turning out to be the limiting factor that limiting factor is actually super serious. The reason it's so serious is the folks with these skill. Nls. If there's not enough of them, they will tend to flock to the places which most sort of welcoming and sets to take them. So when you've got a huge undersupply of a experts in the United States, those that do remain are going to be fought over by very deep pocketed strong internet companies who providing important services. But suddenly you start to see other organizations like even things like NASA war, Veterans, Administration or construction companies. All these other things. Absolutely. Right now need to bring in advanced technology really just being blocked because they come find care. Yeah. So my concern started to be that the biggest problem on need in the future is going to be wet. You get this the supply computer sciences from here with Andrew Moore. He's the dean of Carnegie Mellon's school of computers. Science. So you went there because there is this. It's it's almost a crisis really in terms of like Google. It's essentially Google and Facebook sucking up and others sucking up all the talent, and we'll get into the diversity of the talent in a minute. But so you went there talking about what Carnegie Mellon's doing most people think of Stanford and some other schools in California, polytech and stuff like that. Talk about how you look at computer education now and, and we'll get into the diversity issue in a minute, but how do you look at where we are as a country? Because I just recently read Mark Zuckerberg and he talked about the problem China, obviously, the worrisome nature of what's going on in China is that there's so many people. They're just pushing out a. i. experts everywhere, sort of look at the state of of where we are with graduates in computer science still heavily in demand. So when I look at the whole map of what's needed there. Two distinct populations of folks. We need to train up the both equally important, but they are different. Okay. Number one, the people who can take existing technology the the the great systems, like tends to flow or AWS web services, or at least things integrate them together to make newports. There's the second group which who are the people who are going to be designing the new things, which eventually replace those current. So there's the folks building the technology that the next twenty years will be based on. And then as those taking the existing technology. So a place the the great computer science places MIT Stamford, Georgia Tech, but plea CMU..
"carnegie mellon" Discussed on Recode Decode
"Network today, I'm delighted to have Andrew more on the podcast. He's the dean of Carnegie Mellon's school of computer science, which was ranked number one in the world by US news and World Report, and he was previously vice president of engineering Google, where he was in charge of Google shopping. Andrew, welcome to Rico, decode. Thanks for. Thank you. So let's talk. I wanna I wanna get your background. I have had various computer scientists on the show, and we're teaching and just like that. And I love to get sort of the academic perspective, but you've been in the in the fray also. So let's give your background where you came from and how you got to Carnegie Mellon, and then we'll talk about what's going on there. I grew up in the seaside town called boom in the south of England, and there in the late eighty s I really got into creating video games like kids at the time right? When studied computer science at Cambridge University and then did a PHD on this big question of it's so hard to program robots to do stuff. Even we make them learn to do it instead, which has been the biggest challenge obviously. So that's why I really fell in love with this question of, to what extent can we help machines improve their performance and performance? I, we'll talk about that later a little bit more. So you did that, but you did you go right to robotics? Where'd you go from there? Subsequently, I, I spent some time at the MIT a up which is super fun working for professor Chris Atkinson there, and I'm totally a math statistics guy, where's he builds real robots. So it was a trial. Fan actually build the physical robots, and frankly, I still suck at that mechanical engineer, exactly, huge respect for that the it's the stuff to do with making things decide what they're going to do next. I'm really interested in anyway. Subsequently I joined Connie Email and really enjoyed sort of helping develop the AI classes that got super into using machine learning, not only for robots, but for manufacturing because there's so much I'm do to improve that. I really enjoyed my time that started to get interested in other big questions around computer science to do with things like, can you detect near-earth objects which potentially dangerous using so fancy algorithms or Nathan Myhrvold thing, but go ahead. Can you get an early warning that this being an ABC borne diseases, heck on a city by noticing that the perhaps the uptick in sales of of medications, following stripe along the city in the direction of the f. low for example, that's that was the cool stuff. Right? Helpful is all around this key thing that if you compress a lot of data machines may be able to see stuff that no individual human could see because we can only sort of ingest a certain amount of beta exactly which is the whole idea behind all this. So you were there Carnegie Mellon, and then you went to Google. Is that the only job you've had like that's not academic, or was it? Yes, yes, I did. Do a spectacularly unsuccessful startup for a while. What was it? That was all his spectacularly unsuccessful startup. They're my favorites machine learning consulting services early. Yeah, we in the one thousand nine thousand nine hundred. We had a flashing neon sign on Craig street near CMU which said, data mining mining, fleshing all the time. Right? Which he never got any will can customers, unfortunately. Yeah. Yeah, yeah. Today wouldn't go over well, go ahead what we loved doing that, but we just didn't figure out how to make money at it, consulting engagements on a playing machine learning and all kinds of places really do. Do that was now the thing. So you went over to Google. How did you get to Google? I was really impressed by the way that things were scaling so much and I, I made the move relatively late. It was in the mid two, thousands, and the fact is I was very, very interested in this question of what can you do with a billions or in some cases, more than billions of obsessions. That's what I entice me..
"carnegie mellon" Discussed on Main Engine Cut Off
"Welcome the main engine cutoff. I'm Anthony clans low and we've got some special guests here. We've got Andy and Mike from cube Rover, I ever Andy. Thanks for Anthony. Thanks so much for having us here today. We're really excited to be on your podcast for the first time and hopefully give it a little bit more information about what moon companies are doing. This is we've vaguely touched about moon stuff, mostly policy related on the podcast in the past, but not a lot about actual payloads going to the moon and things like that. So I'm pretty excited to dig into it. So before we get into the technology stuff, that design of cube Rover itself, all that fun stuff. I would really like to hear the roles that you have on the team, and you know what you're working on day to day. So Mike, you want to start. Yeah, sure. So I'm the president of the cube Rover division here at Astra botic, and we're soon spinning out as its own company cube Rover, and I'm the principal investigator on NASA contract under which the key. Rover is being developed in collaboration with Carnegie Mellon University, and that's the small business innovation. What is it was the are. They are. Yeah, I forget what the are is the type of contract used to develop technology focused Pacific toward small like Astra body, and that was a twenty seventeen award. But that came pretty recently. Is that right? Yeah, there is these these phases. So we had a phase one, and then we re proposed for phase two for more dollars and much more. Exciting and bigger development on the project. Leading to that point is kind of interesting because you mentioned that cube Rovers soon to roll off on its own, and you've got these cube Rover, small awards coming from NASA now. So it's seems like an interesting history there. I'm curious to find out how did the project itself start and maybe now the full company, like what is the origin story there. Yeah, I can. I can take that Asser botic and Carnegie Mellon have been working together for ten years since the astronautics inception. And we've developed a lot of Rovers over those years. CMU students and professors have developed their own plant a Rovers over the years all mostly targeting the moon, and it's only been recently that our government has switched its focus, much more towards the moon and started investing in development of technologies, including Rovers to go towards the moon. So what we do is heavy Klabin with with seamew students to develop this Rover is very challenging because of the small size that we're developing. But we have a lot of heritage and learning over the years to lean upon. So what phase would you say cube Rover, both the actual technology, but the company is in at this point in time? Yes. So we're, we're pretty far along with technical progress. We're at what you'd call tier all four going into KIRO five technical technology readiness level. So we have a prototype working driving around in the lab. What we're doing right now is we're trying to build that up to be a flight qualified Rover. So by the end of two thousand or by the middle of two thousand and twenty, we're aiming to have this Rover be ready for flight tested for all space, environmental testing, and ready to send off to NASA if they decide to file it on the company. And we're incorporating right now, we're looking at our location which I can't say where we'll be incorporating, but it will be very soon and we're hoping to have more information on that by the end of the summer. But yeah, we're actually on that note. We're going to be hiring four qb Rover. In probably November or December timeframe. So check out our website for updates on that. Could you talk at all about the decision to roll out as its own standalone enterprise of sorts? Because it seems from from someone like me who's a little bit outside looking in, it seems like a good fit with what Astra botic is offering that you could extend these services to not just Lander, but then have some ability on the surface. So it seems like a good fit there. It is a good fit. It really does align well with us.
"carnegie mellon" Discussed on Grumpy Old Geeks
"That's where they got all the people from carnegie mellon yeah when they when they pulled them away it's looking more and more like oprah's just gonna pull out of the the rnd on this and they'll probably just rent a fleet from tesla or waymo or waymo after you know after that whole thing happen with knowing well we had tech we couldn't make it work so i guess we'll just lease your car's maybe that'll be yeah we'll see how that that pans out because i guess they're really trying to make nice now this new ceo from peres really trying to make nice with google because you know they need them there's yeah there's no way they're going to get you somebody they're not gonna do it by themselves anymore that i mean that's that's become clear so yeah it's either going to be you know waymo tessler gm i this is like the top ones that are going to going to make it i think but we'll see how that plays out but yeah it sucks for these people than in pittsburgh for sure but it doesn't say that they're actually laying off any of the engineers yet so no so far is just the people that were sitting in the cars so yeah tough job yeah now this one this one got my goat i wanted this last week but you weren't around so i waited on it court rules copying photos found on the internet is fair use everybody is pissed off about this one well basically it just says there's no such thing as copyright then like we don't own anything like nobody owns a damn thing and if you put it up on the internet it's now a free for all that's kind of it so how much money to pinterest pour into this thing yes seriously you know their lobbyists were in the back i five in each other when this game down no shit basically gives them a business model which has been lacking for the inception of interest yeah virginia federal court has made a decision the photographers won't be happy to hear no shit the court ruled that finding a photo on the internet and then using it without permission on a commercial website can be considered fair use how okay but how does this not then slippery slope to well i found music on the internet it's fair use i found video on the internet it's fair use i don't understand any differentiation between a photo or any other version of media if you find it on the this to me this ruling will obvious is going after the supreme court and get shut down there's no way we can have diety with this rule in place we can't yet i mean it's a seven page ruling you can go read it if you want the link will be in the show notes it said the use was transformative noncommercial which if you're putting a photograph on a loaded commercial website that's bullshit the use was in good faith oh no god no it wasn't good faith they stole it fuckers in the use was of a factual photo and instead of creative photos in the use was previously published photo well anything on the internet is previously published photo at that point the use was only a crop rather than the whole in the houston hurt the potential market for that photograph well okay there you go there's your loophole but i mean you now you're now you're in the opposite case you're you're you're guilty until proven innocent because now you have to prove that you could have made money off that photo and how do you do that and how do you do that and how are you going to take somebody to court every single time that they use your photo that's ridiculous yeah it's it's not good it's not known so this is going to have to get overturned it must it must which means it probably won't that's good that's a good point welcome to our society can this awesome isn't it this.
"carnegie mellon" Discussed on Part-Time Genius
"Skyrockets and if you can't pay that they get your car or take away your license which then makes it harder to get to your job to pay that ticket and that's what she'd been driving a ton of bankruptcies in the city i mean i guess if you're desperate and you're looking for a fast fix for your financial troubles like playing the lottery seems like the quickest way out at least that's how i think about it but it sounds like you're saying it's more than that like these people are actually being misled or maybe do or something and you know this actually goes back to two thousand eight study like the number of studies i'm quoting today on all these studies i come came with them but this was a group of behavioral economists there at carnegie mellon and so what they were looking to is is is it really like the reason why poor people are so much more likely to play the lottery than those who are better off and what they found was that a lot of the desire that drives these low income players isn't so much from being poor as it is from feeling poor so here's how the study's authors broke down their findings it says in experiment one participants were more likely to purchase lottery tickets when they were primed to perceive that their own income was low relative to the implicit standard and then an experiment to participants purchase more lottery tickets when they considered non lottery situations in which rich people or poor people receive advantages implicitly highlighting the fact that everyone has an equal chance of winning the lottery so basically when you make people feel poor they played the lottery more and when they feel like it's equal like that then they wanna put money on it because they feel like it's a fair game i mean it's kind of heartbreaking when you think about how low their chances of winning actually are though right.
"carnegie mellon" Discussed on Science Friday
"Stop moving she's like losing just like as an exactly exactly like us thank you this is fascinating thing we wish we had more time thank you both for joining us not only is an artist and obama that nvidia research fellow carnegie mellon at the frank ricci studio for creative inquiry henny nonni is a psychologist and assistant professor in the robotics institute at carnegie mellon and director of the human and robot partners lab thank you you can see videos and photos of my guests throw botox work on our website at science friday dot com slash robotics picking us to the break our musical guests for the night pittsburgh's very own townspeople after the break we'll talk about computers that compose their own music and we'll let you hear a pop song written by eight is they would do us and can barely see old bed rooms i am the last son of to smile this is science friday replayed coming to you from the carnegie library music hall in pittsburgh yes and the theme of our show is no assembly required and we've been talking about robots in a i but we have to remember that that behind every robot as we've been talking about it is a piece of software that runs it every row by technology we've been talking about and you'll see tonight is run on some sort of bits and bytes with one exception and that's what we're going to talk about now different kind of robotics my next guest programs the stuff we wear stuff we sit on and now with software she uses bacteria and chemistry and if you wonder what that's about that's what we're going to talk about now she takes actual biological makeup of something let's chair for example and programs that do something it didn't do before like if you like stuff she programs a chair to put itself together.
"carnegie mellon" Discussed on WBT Charlotte News Talk
"The players over the owners will also talk about the new ejection rule for hits that are deemed to be unnecessary they do this in college so under that new rule if an official calls a hit unnecessary they can go review it an object the player on the spot as opposed to you know finding the player later on in the week and suspending him then they could do it right there during the game the owners will decide the location of the next couple of nfl drafts nashville denver vegas cleveland kansas city have all been mentioned and of course they will decide a new owner for the carolina panthers david taffer who has already been vetted by the league is minority owner for the pittsburgh steelers should get the official vote sometime tomorrow making him absolutely the new owner of the carolina panthers he spoke yesterday to the graduates at carnegie mellon university and in his own words said it's been a hell of a week he has just bought the carolina panthers for two point two seven five billion dollars his speech the paper said was emotional at times he shared stories from his past and life lessons that helped pave his success he talked a little bit about his dad said that his dad had work sixty hours a week to make ends meet in highschool tapper was a short order cook at a deli sold knives door to door and worked at a bakery he told the grads mellon yesterday a kid from the streets of san francisco had to work his way through college at pitt and graduate school at cmu just got an honorary degree and is giving a commencement speech at this university he said a kid who couldn't afford to go to an nfl game well into his twenties is now on the verge of getting nfl approval to buy the carolina panthers not too shabby he said he said his greatest accomplishment hasn't been heading appaloosas management his miami based hedge fund it hasn't been his philanthropy or his nfl ownership it's being a better father to his three kids than his physically abusive father was to him and that's when he got motion all he said in.
"carnegie mellon" Discussed on WAFS Biz 1190
"Apartment investor mastery dot com got questions from me want to pick up a copy of the think realty magazine that you can find that the 84 lumber stores orban's noble you know how to do it just gone over there pick it up and you know what it's called conveniently titled think realty magazine so make sure you do that got questions from me abdelhakam hit me up an instagram because i am always answering your questions in its two on instagram all the time an avi call har so let's go and get brad online brad you with me by an area i can let's go get started man i have one question that i generally asked to all my guests before we get the show on the road and that is you can be as dramatic as he walked by the way are you in a real estate of might am i in real estate of mine is that be question bats the question that's the million dollar question for you brand absolutely i mean i got into real estate sixteen years ago and that's been my pashtoon you know both as an investor and as an educator and i mean it's done so many things for my life that that um you know an mba and eventually tear into your aid never never do for may so absolutely that's awesome degree do have you and i have a lot more in common than i thought i'm in the juniors will yeah i graduated from carnegie mellon university back in 1989 uh is in chemical engineering nice that's generally i think i would say that's the hardest engineering degree next two electrical engineering which was my degree from the university of michigan and i think that is so you understand logic you understand flow and i'm really started to have this conversation with you because it's the numbers it's the emotion behind the deal and of course it's getting the deal done so let's start off by introducing a little bit about who you are and what you do yeah.
"carnegie mellon" Discussed on Bloomberg Radio New York
"Of shoes wings to you get a ribbon for having a chartered financial analyst tom ones other did you do i get the world ribbon because i have that i mean what was it like getting a cfa under the guise of military education wealth it was very interesting i'd gone in that long period of time between captain and major after you've commanded you have a significant amount of time i went back and taught of west point frankly my alma mater harvard business will i felt did not heavy strong finance program so so i excellent fiddle funeral now i actually want to learn more about the tough part of finance which was going through a revolution in the 80s as you remember the rocket scientists were coming out the derivatives were being calculating harvard we didn't that's tough to children's occasional up the charles river that maybe a lung no that's exactly right mit and sloan really got hard on that as did the warden in chowdhury carnegie mellon exactly and and being an engineer by background from this point i really wanted to learn the math and god that's my home a mater but that was general management and i was teaching finance what i do felt it was important to know much more the my students will mark tracy patrick kenneth koma's cfa let me i'm going to get to the gossip i can't think of any one more qualified to pick up the pieces of the white house press room than you we have a president who has an affinity for people with ego seasons stars on their shoulders we've had a donnybrook would ever anybody's politics within the press room have you been contacted by general kelly to take over the things you are experienced at which is communicating with the press in the public well i haven't i've got the the grace respect for john kelly serbs sidebyside with them before but i think he said if there are a lot of generals there i spent a year as the spokesman in iraq and that's enough for me see that cold silenced days why did you pick a after righteous destroyed the countries i would it it get your sense of of of how cohesive you think that the national security team is within the the.
"carnegie mellon" Discussed on Chips with Everything
"The guardian when we talk about online society and community issues of power dynamics come up a lot who wields power in social media discourse who feel safe online and who doesn't who feels represented in who doesn't you've probably heard something about manolis fair before you know internet communities where guys go to vent their anger or their fear about women and particularly about feminism of course misogyny itself is nothing new but some of these communities can be pretty toxic even hateful and they've raised a lot of questions for online society i'm meal exander and welcome to chips with everything from the guardian our guest today is angela wash co a game developer writer a teacher at carnegiemellon university angela has been doing both art and research about the self in virtual space for some time now up particularly when it comes to issues of gender and she's even done some deep dives into the manas fear so i'm antle wash co and i'm an artist of video game developer and a writer um i also teacher carnegie mellon university and pittsburgh a shorter transition from during a lot of performances inside massively multiplayer online role playing games most frequently inside world of warcraft i was operating as the council on gender sensitivity m behavioral awareness in world of warcraft for four years um and more recently i've been interested in actually making some of my own games recently i've been working on a a longterm research project.
"carnegie mellon" Discussed on TechStuff
"This actually dates to the 1980s in a computer lab at carnegie mellon university it's amazing that we can actually tracked down the birth of the moada khan as it is used today often these sort of things end up being buried in legend in lower and we never really understand where something came from it just we can can of point when it became popular but anything before that tends to be a mystery not so in this case because things that existed online have a habit of sticking around even if that online was just online in the case of a of a local network and not the internet at large because we're talking about 1982 most universities didn't have access to the internet carnegie mellon probably in exception to that but this was not a internet mean this was very local on a bulletin board system so we can pinpoint the date of a motaqawi creation to september nineteenth nineteen eighty two so here's what was happening computer science students and other students who had become interested in computers uh were using a school electron bullet electronic bulletin board system or be be asked to post messages to one another so this was a predecessor to the news groups and forums that you would find over the internet but this would be a decade before the world wide web ever existed to access abebe s typically you would use a dial up modem and you would call a phone number that would be connected to a specific computer that hosted the bulletin boards system uh some of these bulletin board systems could only have one connection at a time so you might try and calling get a busy signal and you'd have to wait and try and call later you also had la bolton board systems that tried to regulate traffic by charging per minute of use and that way you could cut down on someone just hogging the bulletin boards system just for him or herself.