ai.bythebay.io: Bradford Cross Interview

you you [Music] you so I just go by the sort of definitions so I just go by the sort of definitions from the academia and you know think from the academia and you know think about AI in terms of general techniques about AI in terms of general techniques and planning and so on that have been and planning and so on that have been around for a long time and of course around for a long time and of course machine learning and all the different machine learning and all the different subfields of the machine learning subfields of the machine learning natural language processing computer natural language processing computer vision different niches like this and vision different niches like this and they sort of just think about the field they sort of just think about the field of AI in terms of all these different of AI in terms of all these different subfields that are part of it so we're subfields that are part of it so we're very successful when we have a large very successful when we have a large amount of data large dent data like a amount of data large dent data like a prediction problems ranking problems prediction problems ranking problems like such a Google or Facebook basis like such a Google or Facebook basis that's us all pretty well solved where that's us all pretty well solved where we need to really do a lot of work of we need to really do a lot of work of courses things like unsupervised courses things like unsupervised learning or week or semi-supervised learning or week or semi-supervised learning it's quite difficult to get learning it's quite difficult to get good results still so that's a big area good results still so that's a big area focused and then also areas where focused and then also areas where there's sparse data cases where you there's sparse data cases where you might have only only ability to predict might have only only ability to predict based on empirical evidence in a certain based on empirical evidence in a certain part of the input space but then some part of the input space but then some other parts maybe you have no data at other parts maybe you have no data at all so I think those are also very all so I think those are also very challenging settings where we have a lot challenging settings where we have a lot more work to do is sort of a big dense more work to do is sort of a big dense data type of approaches are pretty well data type of approaches are pretty well solved my key focus is really on applications my key focus is really on applications of machine learning in huge and of machine learning in huge and interesting new deal means outside of interesting new deal means outside of tech you know I feel like a lot of us tech you know I feel like a lot of us have been around the tech world and in have been around the tech world and in the consumer Internet giants like Google the consumer Internet giants like Google Facebook Amazon etc working on these Facebook Amazon etc working on these machines earning problems for 15 years machines earning problems for 15 years or so and the kind of firepower that or so and the kind of firepower that we're bringing hasn't seen as much we're bringing hasn't seen as much action outside of these giant tech action outside of these giant tech companies in these other totally companies in these other totally different domains so my personal focus different domains so my personal focus is on trying to find those business is on trying to find those business problems and interesting technical problems and interesting technical problems that underlie completely new problems that underlie completely new areas where you don't see a lot of the areas where you don't see a lot of the technical firepower in Silicon Valley technical firepower in Silicon Valley currently pointing for me it's it's currently pointing for me it's it's great to be able to come out to these great to be able to come out to these these events right now having gathered these events right now having gathered an experience of doing this stuff for an experience of doing this stuff for such a long period of time and to be such a long period of time and to be able to just share some of those able to just share some of those insights that I know you know maybe not insights that I know you know maybe not everyone's paying attention to but a few everyone's paying attention to but a few people might avoid you know some huge people might avoid you know some huge stumbling blocks or mistakes or failures stumbling blocks or mistakes or failures or waste of time and money and energy or waste of time and money and energy and so that's awesome right I really and so that's awesome right I really enjoyed just being able to come out and enjoyed just being able to come out and share some of the experience viewpoint share some of the experience viewpoint of someone's doing this for quite some of someone's doing this for quite some time and have people come up afterwards time and have people come up afterwards and ask different questions which are and ask different questions which are smart questions and obviously they're smart questions and obviously they're thinking about some of these issues and thinking about some of these issues and trying to avoid some of the same sort of trying to avoid some of the same sort of roadblocks that I am mentioning so for roadblocks that I am mentioning so for me that's great just to be able to share me that's great just to be able to share that and help other people avoid some that and help other people avoid some pain right suppose I think that's that's pain right suppose I think that's that's the thing for me the thing for me you you [Music] you