Curvature-Balanced Feature Manifold Learning for Long-Tailed Classification

by   Yanbiao Ma, et al.

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 systematically propose a series of geometric measurements for perceptual manifolds in deep neural networks, and then 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. Therefore, 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. The code and model will be made public.


page 1

page 2

page 3

page 4


Delving into Semantic Scale Imbalance

Model bias triggered by long-tailed data has been widely studied. Howeve...

Feature-Balanced Loss for Long-Tailed Visual Recognition

Deep neural networks frequently suffer from performance degradation when...

Learning from Long-Tailed Noisy Data with Sample Selection and Balanced Loss

The success of deep learning depends on large-scale and well-curated tra...

Memorization Through the Lens of Curvature of Loss Function Around Samples

Neural networks are overparametrized and easily overfit the datasets the...

Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustment

Long-tailed data is still a big challenge for deep neural networks, even...

Curvature-based Comparison of Two Neural Networks

In this paper we show the similarities and differences of two deep neura...

Emergence of Separable Manifolds in Deep Language Representations

Deep neural networks (DNNs) have shown much empirical success in solving...

Please sign up or login with your details

Forgot password? Click here to reset