Debugging Machine Learning Models

By Gabriel Cadamuro, Ran Gilad-Bachrach, Jerry Zhu. 2016

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Yoshua Bengio
2. Percy Liang
3. Been Kim
4. Mohit Bansal
5. Bo Li
6. Quoc V. Le
7. Nicholas Carlini
8. Pang Wei Koh

Citations in the corpus

• Sanity Checks for Saliency Maps (2018, background): “Explanations could potentially help satisfy regulatory requirements [3], help practitioners debug their model [4, 5], and perhaps, reveal bias or other unintended effects learned by a model [6, 7].”
• Understanding Black-box Predictions via Influence Functions (2017, result): “In contrast to training-set attacks, Cadamuro et al. (2016) consider the task of taking an incorrect test prediction and finding a small subset of training data such that changing the labels on this subset makes the prediction correct.”
• Debugging Tests for Model Explanations (2020, background): “Diagnosing and fixing model errors–model debugging–remains a longstanding machine learning challenge [12, 14–17, 55, 73].”
• An Overview of Machine Teaching (2018, background): “This ranges from angelic (optimized, personalized education; improving cognitive models; fast classifier training; debugging machine learners [7, 17,46], etc.”