Machine learning pitfalls

Tech Valley Machine Learning, Data Science, and AI
Tech Valley Machine Learning, Data Science, and AI
Public group
Location image of event venue


• What we'll do
Building off David Martuscello's intro talk at the last meetup, Dan Elton will present some pitfalls one can encounter with machine learning. It is important not to get caught up in the hype surrounding machine learning and maintain a sense of what ML really is and what its limitations are. The pitfalls to be discussed are:
-- overfitting & underfitting
-- not cleaning & normalizing your data
-- trying to use machine learning for extrapolation
-- biased data sampling
-- not doing hyperparameter optimization carefully
-- not comparing your results to simple baselines
-- trying to infer causal mechanisms from machine learning models
-- sloppy ranking of feature importance

• What to bring

• Important to know
Target audience: everyone is welcome; the talk will assume basic knowledge of machine learning, but I will also try to keep it light and fun so everyone can get something out of it. We thank Matthew Lean and OneHudson Ventures ( for letting us use the Estuari coworking space.