[2603.02280] Temporal Imbalance of Positive and Negative Supervision in Class-Incremental Learning

Computer Science > Machine Learning arXiv:2603.02280 (cs) [Submitted on 2 Mar 2026] Title:Temporal Imbalance of Positive and Negative Supervision in Class-Incremental Learning Authors:Jinge Ma, Fengqing Zhu View a PDF of the paper titled Temporal Imbalance of Positive and Negative Supervision in Class-Incremental Learning, by Jinge Ma and 1 other authors View PDF HTML (experimental) Abstract:With the widespread adoption of deep learning in visual tasks, Class-Incremental Learning (CIL) has become an important paradigm for handling dynamically evolving data distributions. However, CIL faces the core challenge of catastrophic forgetting, often manifested as a prediction bias toward new classes. Existing methods mainly attribute this bias to intra-task class imbalance and focus on corrections at the classifier head. In this paper, we highlight an overlooked factor -- temporal imbalance -- as a key cause of this bias. Earlier classes receive stronger negative supervision toward the end of training, leading to asymmetric precision and recall. We establish a temporal supervision model, formally define temporal imbalance, and propose Temporal-Adjusted Loss (TAL), which uses a temporal decay kernel to construct a supervision strength vector and dynamically reweight the negative supervision in cross-entropy loss. Theoretical analysis shows that TAL degenerates to standard cross-entropy under balanced conditions and effectively mitigates prediction bias under imbalance. Extensive ex...