ai.bythebay.io: Arno Candel Interview

you you [Music] you [Music] ai is of course a field in computer ai is of course a field in computer science where people try to make science where people try to make autonomous agents smart and behave like autonomous agents smart and behave like humans but what I really means for most humans but what I really means for most businesses that's who we are targeting businesses that's who we are targeting is basically make better decisions there is basically make better decisions there currently is based on rules based currently is based on rules based distance so where some human has distance so where some human has programmed something like if H is more programmed something like if H is more than 5 and income less than 50,000 then than 5 and income less than 50,000 then deny or something and these rules deny or something and these rules getting out of placed by smarter getting out of placed by smarter machines that automatically generate machines that automatically generate those rules and you can have a billion those rules and you can have a billion rules that are auto-generated and they rules that are auto-generated and they might be more complicated than those if might be more complicated than those if else decisions but those complicated else decisions but those complicated decisions basically make better choices decisions basically make better choices than humans whether that's playing a than humans whether that's playing a computer game or making business computer game or making business transactional decisions for translating transactional decisions for translating languages that's basically AI for languages that's basically AI for enterprise so AI has made huge strides in so AI has made huge strides in translating languages and looking at translating languages and looking at images understanding what's in the images understanding what's in the images understanding sound these images understanding sound these unstructured data sets but but it's unstructured data sets but but it's still not good at is in figuring out still not good at is in figuring out what the humans actually wants so for what the humans actually wants so for example if you have a data set of example if you have a data set of transactions everybody who's ever swipe transactions everybody who's ever swipe the credit card in the last two years the credit card in the last two years now you could ask a lot of questions and now you could ask a lot of questions and AI has no clue what questions to ask or AI has no clue what questions to ask or has no clue about who to deny the credit has no clue about who to deny the credit for example or not or who to go after for example or not or who to go after and say I can double your credit or I and say I can double your credit or I can give you a better account or can give you a better account or something like that ai doesn't know that something like that ai doesn't know that on its own right so it still needs on its own right so it still needs humans to make up those questions and humans to make up those questions and then use the answers to decide what then use the answers to decide what kitty so the translations of the kitty so the translations of the business is still lacking so at h2o I business is still lacking so at h2o I and me as a CTO I'm varied and I'm and me as a CTO I'm varied and I'm focusing on making the customer happy focusing on making the customer happy write a customer says mella has this write a customer says mella has this insurance problem I need to figure out insurance problem I need to figure out what my mother ate should be fully what my mother ate should be fully insurance and how do i best model that insurance and how do i best model that and so my focus is on making better and so my focus is on making better algorithms better data science workflow algorithms better data science workflow so that the arbitrary people in the so that the arbitrary people in the market out there can be more productive market out there can be more productive whether that's a data scientist or even whether that's a data scientist or even the business analyst or the super geeky the business analyst or the super geeky Cargill master I want to please them all Cargill master I want to please them all so I stay up at night trying figuring so I stay up at night trying figuring out places where i can improve this kind out places where i can improve this kind of virtual experience so that everybody of virtual experience so that everybody can be more productive I like all AI conferences because it's I like all AI conferences because it's such a thriving feels right now and such a thriving feels right now and every week there's a new paper on every week there's a new paper on something new and exciting and I like to something new and exciting and I like to see how people are applying these new see how people are applying these new methods to a new use cases and pushing methods to a new use cases and pushing forward the envelope of AI forward the envelope of AI [Music] you you