Developing a ML model
To a person new to it, Machine Learning can feel more like art than science. I’m adding layers and hidden units but nothing seems to increase the performance of my model. I just paid half my paycheck to NVIDIA for a new GPU, but my model training is not much faster than before. In this session, we will try to trade in some art to get back some science.
We will look at some options to power our applications with Machine Learning without the challenges of building our own model. We will then look at leveraging the work other people have done and published to simplify our model building. Finally, for those of us who are not content with anything less than building our own very complex models with a lot of data, we will look at strategies around distributed Machine Learning.
-
Exploring StackOverflow DataEvelina GabasovaWednesday Apr 25, 11:15
-
Developing a ML modelKevin TsaiWednesday Apr 25, 13:00
-
Production Model DeploymentJuliet HouglandWednesday Apr 25, 15:15
-
Life and Death Decisions: Testing Data SciencePhil WinderWednesday Apr 25, 14:00
-
Relating to Machine LearningStefan Veis PennerupWednesday Apr 25, 10:15
-
Delivering AI on Code: Live Demo of source{d}Francesc CampoyWednesday Apr 25, 16:15