Feedback-guided Data Synthesis for Imbalanced Classification

ICLR 2024 Workshop DMLR Submission24 Authors

Keywords: Dataset Augmentation, Synthetic generation, Diffusion Models, Imbalanced Classification

TL;DR: We introduce a framework for synthetic data generation with classifier-generator feedback mechanism that improves the performance on imbalanced classification problems.

Abstract: Current status quo in machine learning is to use static datasets of real images for training, which often come from long-tailed distributions. With the recent advances in generative models, researchers have started augmenting these static datasets with synthetic data, reporting moderate performance improvements on classification tasks. We hypothesize that these performance gains are limited by the lack of feedback from the classifier to the generative model, which would promote the usefulness of the generated samples to improve the classifier’s performance. In this work, we introduce a framework for augmenting static datasets with useful synthetic samples, which leverages one-shot feedback from the classifier to drive the sampling of the generative model. In order for the framework to be effective, we find that the samples must be close to the support of the real data of the task at hand, and be sufficiently diverse. We validate three feedback criteria on a long-tailed dataset (ImageNet-LT) as well as a group-imbalanced dataset (NICO++). On ImageNet-LT, we achieve state-of-the-art results, with over 4% improvement on underrepresented classes while being twice efficient in terms of the number of generated synthetic samples. NICO++ also enjoys marked boosts of over 5% in worst group accuracy. With these results, our framework paves the path towards effectively leveraging state-of-the-art text-to-image models as data sources that can be queried to improve downstream applications.

Primary Subject Area: Other

Paper Type: Research paper: up to 8 pages

Participation Mode: In-person

Confirmation: I have read and agree with the workshop's policy on behalf of myself and my co-authors.

Submission Number: 24