Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

regularization-based CIL

Regularization-based class-incremental learning is an approach in continual machine learning where a model learns to recognize new object classes in sequential stages while preventing catastrophic forgetting of earlier classes through penalty terms added to the loss function. Rather than relying on storing past training samples or continually expanding the neural network architecture, these methods preserve prior knowledge by constraining updates to parameters identified as important for previous tasks or by employing knowledge distillation to maintain consistent intermediate and output representations. This strategy allows the model to incrementally adapt its decision boundaries for newly arriving classes while retaining its performance on previously mastered categories in a memory-efficient and privacy-preserving manner.

1 item

ACIL: Analytic Class-Incremental Learning with Absolute Memorization and Privacy Protection

ACIL: Analytic Class-Incremental Learning with Absolute Memorization and Privacy Protection

Huiping Zhuang, Zhenyu Weng, Hongxin Wei, Renchunzi Xie, Kar-Ann Toh, Zhiping Lin

OrganizationsNanyang Technological UniversitySouth China University of TechnologyYonsei University

Why you should read this

Proposes an analytic class-incremental learning framework that mathematically matches the performance of joint training without storing historical exemplar data, eliminating catastrophic forgetting while protecting data privacy across multi-phase learning tasks.

Class-incremental learning (CIL) learns a classification model with training data of different classes arising progressively. Existing CIL either suffers from serious accuracy loss due to catastrophic forgetting, or invades data privacy by revisiting used exemplars. Inspired by linear learning formulations, we propose an analytic class-incremental learning (ACIL) with absolute memorization of past knowledge while avoiding breaching of data privacy (i.e., without storing historical data). The absolute memorization is demonstrated in the sense that class-incremental learning using ACIL given present data would give identical results to that from its joint-learning counterpart which consumes both present and historical samples. This equality is theoretically validated. Data privacy is ensured since no historical data are involved during the learning process. Empirical validations demonstrate ACIL's competitive accuracy performance with near-identical results for various incremental task settings (e.g., 5-50 phases). This also allows ACIL to outperform the state-of-the-art methods for large-phase scenarios (e.g., 25 and 50 phases).

Added

2026-09-26