Lineage-based pruning of ML tasks

Detect user-provided pruning conditions via lineage traces and gracefully abort or move to the next iteration. Examples [1] include depth of a decision tree too large, loss not decreasing, class imbalance, not normalized data, custom pruning conditions.

[1] Nathalie Rauschmayr et al., 2021. AMAZON SAGEMAKER DEBUGGER: A SYSTEM FOR REAL-TIME INSIGHTS INTO MACHINE LEARNING MODEL TRAINING. MLSys.