About the talk
Welcome to Addison's TV here from the AI Summit San Francisco 2019. We very few to be doing today by each division. How are you? Yes, that's what I work for a company called axles and Berg equities and we do asset management. So we managed assets over the world and title isn't the director of currency Investment Management. So how many students are from the from the ingestion to go to the models with my hosta like Leverage AI Technologies plus a general introduction
about what he means they've been doing it for more than 20 years. That was doing all that work. Should I cook a 25 years ago published the book on these 20 20 years ago. I get some kind of perspective and keeps on doing it and if I Reply to say that what was a is Mona volution the Nueva Lucian. So it's a continuation of a set of tools that makes it possible for us to whoever Weds companies in to trade companies near me that's interesting and the amount of data and the cost of a
computation of power. It's interesting to us to question what has not changed and it's interesting that the Yahoo isn't themselves. I'm not changed very much for everything. He's called the back propagation. I'll go in and then there was published for the first time about 20 years ago that still being used now for muscle machine. so in that sense, it's interesting that stay the same and now they're being used for a short text or between traditional pattern recognition in General Tso date
25 years ago would take me two seconds or less to milliseconds now to find the same solution what have the major strides being in the last couple of years for the for when I say we always use algorithms to find values in companies on the guy, you know the value so machine learning that says no networks is a Taser International Revolution the way we do so we have now this is us for years. So we use no network. She's throughout to invest so that as being a changed and also we use more and more natural language processing. So
and we analyzed running calls and texts and we'll use one will send him and do that. Also being generated from text. So I enter A lot of conversation here around kind of the trains going to see in the next year or two years and he's like that. What what do you kind of predict that we received was within machine that it says about the application. So how it's going to go to be be evolving than a minute. There is a lot of work to be done anyway to understanding language. I think you sent what we do now and if he's almost currently condition the text which is
very different from the people really understanding the main and we have been doing what we call the expert system which is in the way taking the you know, that your people and pulling into software so that we can use neural networks on all natural language processing. It's always using what people understand and putting back into the machine. So I think that's possible evolved natural language processing is going to be Solving a Lockton understand language that's going to change. I think I hope
in the in the field itself in all domain this trend of AI machine learning natural language processing. That's a major Trend muscle type of data related to the environment is also going to change. So we all know climate change is a big thing. I know he's now a lot of convergence between using the techniques for an auto put for you is also include on the environment. So we also converging that direction.
Buy this talk
Access to all the recordings of the event
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.