Balancing the Causal Effects in Class-Incremental Learning

arXiv:2402.10063v1 Announce Type: new
Abstract: Class-Incremental Learning (CIL) is a practical and challenging problem for achieving general artificial intelligence. Recently, Pre-Trained Models (PTMs) have led to breakthroughs in both visual and natural language processing tasks. Despite recent studies showing PTMs' potential ability to learn sequentially, a plethora of work indicates the necessity of alleviating the catastrophic forgetting of PTMs. Through a pilot study and a causal analysis of CIL, we reveal that the crux lies in the imbalanced causal effects …

abstract and natural language processing artificial artificial intelligence arxiv class cs.lg effects general general artificial intelligence incremental intelligence language language processing learn natural natural language natural language processing practical pre-trained models processing studies tasks type visual work