DP-CRE: Continual Relation Extraction via Decoupled Contrastive Learning and Memory Structure Preservation

arXiv:2403.02718v1 Announce Type: new
Abstract: Continuous Relation Extraction (CRE) aims to incrementally learn relation knowledge from a non-stationary stream of data. Since the introduction of new relational tasks can overshadow previously learned information, catastrophic forgetting becomes a significant challenge in this domain. Current replay-based training paradigms prioritize all data uniformly and train memory samples through multiple rounds, which would result in overfitting old tasks and pronounced bias towards new tasks because of the imbalances of the replay set. To handle …

abstract arxiv catastrophic forgetting challenge continual continuous cs.cl current data domain extraction information introduction knowledge learn memory preservation relational tasks training type via