2022/06/25 by Tina Behnia, Ke Wang, Behnia, Tina +3
Chemistry · Computer Science · Decision Sciences · #Advanced Statistical Process Monitoring #FOS: Computer and information sciences #Imbalanced Data Classification Techniques #Machine Learning (cs.LG) #Spectroscopy and Chemometric Analyses
paper · pdf · doi:10.48550/arxiv.2206.12739
openalex publication_date 2022/06/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Overparameterized models fail to generalize well in the presence of data imbalance even when combined with traditional techniques for mitigating imbalances. This paper focuses on imbalanced classification datasets, in which a small subset of the population -- a minority -- may contain features that correlate spuriously with the class label. For a parametric family of cross-entropy loss modifications and a representative Gaussian mixture model, we derive non-asymptotic generalization bounds on the worst-group error that shed light on the role of different hyper-parameters. Specifically, we prove that, when appropriately tuned, the recently proposed VS-loss learns a model that is fair towards minorities even when spurious features are strong. On the other hand, alternative heuristics, such as the weighted CE and the LA-loss, can fail dramatically. Compared to previous works, our bounds hold for more general models, they are non-asymptotic, and, they apply even at scenarios of extreme imbalance.