Can You Rely on Your Model Evaluation? Improving Model Evaluation with Synthetic Test Data
This addresses the challenge of ensuring fairness and reliability in real-world applications for machine learning practitioners, though it is incremental as it builds on existing generative modeling techniques.
The paper tackles the problem of evaluating machine learning models on diverse and underrepresented subgroups by introducing 3S Testing, a deep generative modeling framework that generates synthetic test data for small subgroups and simulates distributional shifts, outperforming traditional baselines in estimating model performance and providing superior coverage intervals.
Evaluating the performance of machine learning models on diverse and underrepresented subgroups is essential for ensuring fairness and reliability in real-world applications. However, accurately assessing model performance becomes challenging due to two main issues: (1) a scarcity of test data, especially for small subgroups, and (2) possible distributional shifts in the model's deployment setting, which may not align with the available test data. In this work, we introduce 3S Testing, a deep generative modeling framework to facilitate model evaluation by generating synthetic test sets for small subgroups and simulating distributional shifts. Our experiments demonstrate that 3S Testing outperforms traditional baselines -- including real test data alone -- in estimating model performance on minority subgroups and under plausible distributional shifts. In addition, 3S offers intervals around its performance estimates, exhibiting superior coverage of the ground truth compared to existing approaches. Overall, these results raise the question of whether we need a paradigm shift away from limited real test data towards synthetic test data.