From Computational to Cognitive Systems - Ryan Steelberg President of Vertione on AI Applications that Mine Unstructured Data - Voicebot Podcast Ep 222 - burst 06

Automatic TRANSCRIPT

Big transition that we've been dealing with your in terms of ai. Machine learning is really transitioning from computational systems in decision making to cognitive right decision making out in insights so Baritone fact became a open platform. Call on an operating system for os. That is really good at ingesting. All different types of data then being able to use our suite of over four hundred fifty different cognitive engines and models to to be able to analyze interrogate and ultimately you know empower our business customers with you know the tools applications in unstructured data so they can advance their business. Why do you have so many machine learning models because every job it potentially warrants a different bespoke recipe of cognition. So if i say for example Where presented with australian english. Right you absolutely potentially want to use a different train model as to glowing. Wash 'em in saint louis true for like for example variances in near field or barfield audio right throwing one transcription engine Tire corpus of audio You're gonna get you will get varying results in. Our job is to be able to analyze the data sets in ben in effect programmatic find in orchestrate. The right hung recipe to get the best results and to be clear. I'll give you some examples. The best results is not always accuracy. Right if you're dealing with so much tonnage of data some projects that we work on. You know they're okay with eighty percent accuracy. Because it's it's there's so much volume. They have to meet a price point right. They can't wait ten dollars or fifteen dollars an hour of content process it so and that's in that thankfully that decision being kind of model agnostic in being able to normalize. The output of these different cognitive classes really was the cornerstones of verizon's

Coming up next