Litcius/Paper detail

Knowledge-Augmented Language Models for Cause-Effect Relation Classification

Pedram Hosseini, David Broniatowski, Mona Diab

202217 citationsDOIOpen Access PDF

Abstract

Previous studies have shown the efficacy of knowledge augmentation methods in pretrained language models. However, these methods behave differently across domains and downstream tasks. In this work, we investigate the augmentation of pretrained language models with knowledge graph data in the causeeffect relation classification and commonsense causal reasoning tasks. After automatically verbalizing triples in ATOMIC 20 20 , a wide coverage commonsense reasoning knowledge graph, we continually pretrain BERT and evaluate the resulting model on cause-effect pair classification and answering commonsense causal reasoning questions. Our results show that a continually pretrained language model augmented with commonsense reasoning knowledge outperforms our baselines on two commonsense causal reasoning benchmarks, COPA and BCOPA-CE, and a Temporal and Causal Reasoning (TCR) dataset, without additional improvement in model architecture or using quality-enhanced data for fine-tuning.

Topics & Concepts

Computer scienceCommonsense reasoningCommonsense knowledgeNatural language processingArtificial intelligenceLanguage modelRelation (database)Causal reasoningQuestion answeringGraphCausal modelModel-based reasoningKnowledge graphMachine learningKnowledge representation and reasoningCognitionData miningPsychologyTheoretical computer sciencePathologyNeuroscienceMedicineTopic ModelingNatural Language Processing Techniques