We all love linear regression for its interpretability: Increase square meters by 1, that leads to rent going up by 8 euros. A human can easily understand why this model made a certain prediction.
Complex machine learning models like tree aggregates or neural networks usually make better predictions, but this comes at a price: it's hard to understand these models.
In this talk, we'll look at a few methods to pry open these models and gain some insights
Specifically, the topics covered are
- What makes a model interpretable?
- Linear models, trees, decision rules
- The SIPA framework (Sampling, Intervention, Prediction, Aggregation) for making models interpretable again
- Model-agnostic methods for interpretability
- Partial dependence plots (PDPs)
- Individual conditional expectations (ICEs)
- Example-based explanations
- The future of machine learning