Continual Learning by Three-Phase Consolidation

arXiv:2403.14679v1 Announce Type: new
Abstract: TPC (Three-Phase Consolidation) is here introduced as a simple but effective approach to continually learn new classes (and/or instances of known classes) while controlling forgetting of previous knowledge. Each experience (a.k.a. task) is learned in three phases characterized by different rules and learning dynamics, aimed at removing the class-bias problem (due to class unbalancing) and limiting gradient-based corrections to prevent forgetting of underrepresented classes. Several experiments on complex datasets demonstrate its accuracy and efficiency advantages …

abstract arxiv bias class consolidation continual cs.cv cs.lg dynamics experience instances knowledge learn rules simple type