Decision Boundary-aware Generation for Long-tailed Learning
For researchers tackling long-tailed learning, this work addresses a previously overlooked side effect of head-to-tail transfer, offering a method to improve classifier decision boundaries.
The paper identifies that head-to-tail transfer in diffusion-based generative augmentation for long-tailed learning causes decision boundary overlap and tail class distribution shift. To address this, they propose the Decision Boundary-aware Generation (DBG) framework, which generates informative near-boundary samples to improve decision space separability, consistently improving tail class and overall accuracy on standard benchmarks.
Long-tailed data bias decision boundaries toward head classes and degrade tail class accuracy. Diffusion-based generative augmentation address this problem by generating additional data, while head-to-tail transfer further mitigate the generator bias inherit from long-tailed dataset. However, we show that while head-to-tail transfer helps balance the decision space of the classifier, it also induces latent non-local feature mixing that entangles inter-class features, causing decision boundary overlap and tail class distribution shift. To address this, we first identify the problem of boundary ambiguity and then propose Decision Boundary-aware Generation (DBG) framework, which promotes near-boundary representation learning by generating informative near-boundary samples. Overall, DBG rebalances the long-tailed dataset while yielding more separable decision space for long-tailed learning. Across standard long-tailed benchmarks, DBG consistently improves tail class and overall accuracy with less inter-class overlap. The code of DBG is available at https://github.com/keepdigitalabc-svg/DBG.