CLASSIC: Continual and Contrastive Learning of Aspect Sentiment Classification Tasks

TL;DR

This paper introduces CLASSIC, a novel contrastive continual learning model for aspect sentiment classification in domain incremental settings, enabling effective knowledge transfer without task identifiers during testing.

Contribution

It proposes a new contrastive continual learning approach tailored for domain incremental aspect sentiment classification, addressing the challenge of task-agnostic testing.

Findings

High effectiveness of CLASSIC in experiments

Successful knowledge transfer across tasks

Elimination of task ID requirement during testing

Abstract

This paper studies continual learning (CL) of a sequence of aspect sentiment classification(ASC) tasks in a particular CL setting called domain incremental learning (DIL). Each task is from a different domain or product. The DIL setting is particularly suited to ASC because in testing the system needs not know the task/domain to which the test data belongs. To our knowledge, this setting has not been studied before for ASC. This paper proposes a novel model called CLASSIC. The key novelty is a contrastive continual learning method that enables both knowledge transfer across tasks and knowledge distillation from old tasks to the new task, which eliminates the need for task ids in testing. Experimental results show the high effectiveness of CLASSIC.