Predicting and Enhancing the Fairness of DNNs with the Curvature of Perceptual Manifolds
Authors:
- Yanbiao MaYMYanbiao Ma
- Licheng JiaoLJLicheng Jiao
- Fang LiuFLFang Liu
- Maoji WenMWMaoji Wen
- Lingling LiLLLingling Li
- Wenping MaWMWenping Ma
- Shuyuan YangSYShuyuan Yang
- Xu LiuXLXu Liu
- Puhua ChenPCPuhua Chen
Preprint
Abstract
To address the challenges of long-tailed classification, researchers have proposed several approaches to reduce model bias, most of which assume that classes with few samples are weak classes. However, recent studies have shown that tail classes are not always hard to learn, and model bias has been observed on sample-balanced datasets, suggesting the existence of other factors that affect model bias. In this work, we first establish a geometric perspective for analyzing model fairness and then systematically propose a series of geometric measurements for perceptual manifolds in deep neural networks. Subsequently, we comprehensively explore the effect of the geometric characteristics of perceptual manifolds on classification difficulty and how learning shapes the geometric characteristics of perceptual manifolds. An unanticipated finding is that the correlation between the class accuracy and the separation degree of perceptual manifolds gradually decreases during training, while the negative correlation with the curvature gradually increases, implying that curvature imbalance leads to model bias.Building upon these observations, we propose curvature regularization to facilitate the model to learn curvature-balanced and flatter perceptual manifolds. Evaluations on multiple long-tailed and non-long-tailed datasets show the excellent performance and exciting generality of our approach, especially in achieving significant performance improvements based on current state-of-the-art techniques. Our work opens up a geometric analysis perspective on model bias and reminds researchers to pay attention to model bias on non-long-tailed and even sample-balanced datasets.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- Geometric Prior Guided Feature Representation Learning for Long-Tailed ClassificationInternational Journal of Computer Vision, 2024
- Class-Difficulty Based Methods for Long-Tailed Visual RecognitionInternational Journal of Computer Vision, 2022
- The Devil Is in Classification: A Simple Framework for Long-Tail Instance Segmentation2020
- A Geometric Understanding of Deep LearningEngineering, 2020
- Separability and geometry of object manifolds in deep neural networksNature Communications, 2020
- Adaptive Cost-Sensitive Online ClassificationIEEE Transactions on Knowledge and Data Engineering, 2018
- Linear readout of object manifoldsPhysical Review E, 2016
- Training cost-sensitive neural networks with methods addressing the class imbalance problemIEEE Transactions on Knowledge and Data Engineering, 2005
- A Multiple Resampling Method for Learning from Imbalanced Data SetsComputational Intelligence, 2004
- A Global Geometric Framework for Nonlinear Dimensionality ReductionScience, 2000