Order-Robust Class Incremental Learning: Graph-Driven Dynamic Similarity Grouping
This addresses a critical but understudied challenge in incremental learning for AI systems that need to adapt to new data over time, offering a novel solution to enhance model stability.
The paper tackles the problem of class order sensitivity in Class Incremental Learning, where model performance varies with the sequence of new classes, especially when they are similar, and proposes a graph-based grouping method that improves robustness and achieves optimal accuracy and anti-forgetting results.
Class Incremental Learning (CIL) aims to enable models to learn new classes sequentially while retaining knowledge of previous ones. Although current methods have alleviated catastrophic forgetting (CF), recent studies highlight that the performance of CIL models is highly sensitive to the order of class arrival, particularly when sequentially introduced classes exhibit high inter-class similarity. To address this critical yet understudied challenge of class order sensitivity, we first extend existing CIL frameworks through theoretical analysis, proving that grouping classes with lower pairwise similarity during incremental phases significantly improves model robustness to order variations. Building on this insight, we propose Graph-Driven Dynamic Similarity Grouping (GDDSG), a novel method that employs graph coloring algorithms to dynamically partition classes into similarity-constrained groups. Each group trains an isolated CIL sub-model and constructs meta-features for class group identification. Experimental results demonstrate that our method effectively addresses the issue of class order sensitivity while achieving optimal performance in both model accuracy and anti-forgetting capability. Our code is available at https://github.com/AIGNLAI/GDDSG.