How do you explain individual predictions in production?
mediumAnswer
- (1) SHAP: computes Shapley values per feature contribution.
- TreeSHAP fast for trees.
- (2) LIME: local linear approximation.
- (3) Integrated gradients / Grad-CAM for deep models.
- (4) Attention weights for transformers (interpretation with caution).
- (5) Feature importance dashboards.
- (6) Counterfactual explanations: 'if X had been Y, prediction would be Z'.
- (7) Store per-prediction explanations for regulatory / customer service.
- Tools: SHAP, Captum, InterpretML.
Check yourself — multiple choice
- Random
- SHAP (Shapley) + LIME (local linear) + integrated gradients / Grad-CAM + attention (careful) + counterfactuals + per-prediction storage; SHAP / Captum / InterpretML
- Not possible
- Just guess
Explainability: SHAP / LIME / IG / counterfactuals; store for regulatory.
#monitoring#safety
Practise MLOps & Data Quality
215 interview questions in this topic.