CALA: A Class-Aware Logit Adapter for Few-Shot Class-Incremental Learning
Authors:
- Chengyan LiuCLChengyan Liu
- Fuyuan HuFHFuyuan Hu
- Tao ZhouTZTao Zhou
Preprint
Abstract
Few-Shot Class-Incremental Learning (FSCIL) defines a practical but challenging task where models are required to continuously learn novel concepts with only a few training samples. Due to data scarcity, existing FSCIL methods resort to training a backbone with abundant base data and then keeping it frozen afterward. However, the above operation often causes the backbone to overfit to base classes while overlooking the novel ones, leading to severe confusion between them. To address this issue, we propose Class-Aware Logit Adapter (CALA). Our method involves a lightweight adapter that learns to rectify biased predictions through a pseudo-incremental learning paradigm. In the real FSCIL process, we use the learned adapter to dynamically generate robust balancing factors. These factors can adjust confused novel instances back to their true label space based on their similarity to base classes. Specifically, when confusion is more likely to occur in novel instances that closely resemble base classes, greater rectification is required. Notably, CALA operates on the classifier level, preserving the original feature space, thus it can be flexibly plugged into most of the existing FSCIL works for improved performance. Experiments on three benchmark datasets consistently validate the effectiveness and flexibility of CALA. Codes will be available upon acceptance.Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- Adjusting Logit in Gaussian Form for Long-Tailed Visual RecognitionIEEE Transactions on Artificial Intelligence, 2024
- Few-Shot Class-Incremental Learning by Sampling Multi-Phase TasksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2023
- Learning to Learn Task-Adaptive Hyperparameters for Few-Shot LearningIEEE Transactions on Pattern Analysis and Machine Intelligence, 2023
- Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration2023
- ImageNet Large Scale Visual Recognition Challenge