Spurious Correlations in Concept Drift: Can Explanatory Interaction Help?
This addresses concept drift detection for ML models in dynamic environments, but it appears incremental as it builds on existing detection methods by incorporating explanations and feedback.
The paper tackles the problem of spurious correlations spoiling concept drift detection in long-running ML models, introducing ebc-exstream, a detector that uses model explanations and human feedback to correct these correlations, with preliminary experiments showing promise in reducing their impact.
Long-running machine learning models face the issue of concept drift (CD), whereby the data distribution changes over time, compromising prediction performance. Updating the model requires detecting drift by monitoring the data and/or the model for unexpected changes. We show that, however, spurious correlations (SCs) can spoil the statistics tracked by detection algorithms. Motivated by this, we introduce ebc-exstream, a novel detector that leverages model explanations to identify potential SCs and human feedback to correct for them. It leverages an entropy-based heuristic to reduce the amount of necessary feedback, cutting annotation costs. Our preliminary experiments on artificially confounded data highlight the promise of ebc-exstream for reducing the impact of SCs on detection.