Timezone: »

 
Poster
No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems
Nimit Sohoni · Jared Dunnmon · Geoffrey Angus · Albert Gu · Christopher Ré

Mon Dec 07 09:00 PM -- 11:00 PM (PST) @ Poster Session 0 #3
in Poster Session 1 »

In real-world classification tasks, each class often comprises multiple finer-grained "subclasses." As the subclass labels are frequently unavailable, models trained using only the coarser-grained class labels often exhibit highly variable performance across different subclasses. This phenomenon, known as hidden stratification, has important consequences for models deployed in safety-critical applications such as medicine. We propose GEORGE, a method to both measure and mitigate hidden stratification even when subclass labels are unknown. We first observe that unlabeled subclasses are often separable in the feature space of deep neural networks, and exploit this fact to estimate subclass labels for the training data via clustering techniques. We then use these approximate subclass labels as a form of noisy supervision in a distributionally robust optimization objective. We theoretically characterize the performance of GEORGE in terms of the worst-case generalization error across any subclass. We empirically validate GEORGE on a mix of real-world and benchmark image classification datasets, and show that our approach boosts worst-case subclass accuracy by up to 22 percentage points compared to standard training techniques, without requiring any prior information about the subclasses.

Author Information

Nimit Sohoni (Stanford University)
Jared Dunnmon (Stanford University)

-- Postdoctoral scholar with Prof. Chris Re at Stanford University -- Foci include multi-modal weak supervision and data augmentation

Geoffrey Angus (Stanford University)
Albert Gu (Stanford)
Christopher Ré (Stanford)

More from the Same Authors