ComFact: A Benchmark for Linking Contextual Commonsense Knowledge
This addresses the challenge of retrieving relevant commonsense knowledge for NLP systems in dialogues and stories, though it is incremental as it builds on existing knowledge graph methods.
The paper tackles the problem of linking contextual commonsense knowledge to narratives by introducing the ComFact benchmark, which contains ~293k annotations across four datasets, and shows that learned models improve F1 by ~34.6% over heuristics and boost downstream task performance by 9.8% on average.
Understanding rich narratives, such as dialogues and stories, often requires natural language processing systems to access relevant knowledge from commonsense knowledge graphs. However, these systems typically retrieve facts from KGs using simple heuristics that disregard the complex challenges of identifying situationally-relevant commonsense knowledge (e.g., contextualization, implicitness, ambiguity). In this work, we propose the new task of commonsense fact linking, where models are given contexts and trained to identify situationally-relevant commonsense knowledge from KGs. Our novel benchmark, ComFact, contains ~293k in-context relevance annotations for commonsense triplets across four stylistically diverse dialogue and storytelling datasets. Experimental results confirm that heuristic fact linking approaches are imprecise knowledge extractors. Learned fact linking models demonstrate across-the-board performance improvements (~34.6% F1) over these heuristics. Furthermore, improved knowledge retrieval yielded average downstream improvements of 9.8% for a dialogue response generation task. However, fact linking models still significantly underperform humans, suggesting our benchmark is a promising testbed for research in commonsense augmentation of NLP systems.