DiscoSense: Commonsense Reasoning with Discourse Connectives

Prajjwal Bhargava and Vincent Ng.
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp. 10295-10310, 2022.

Click here for the PDF version.

Abstract

We present DiscoSense, a benchmark for commonsense reasoning via understanding a wide variety of discourse connectives. We generate compelling distractors in DiscoSense using Conditional Adversarial Filtering, an extension of Adversarial Filtering that employs conditional generation. We show that state-of-the-art pre-trained language models struggle to perform well on DiscoSense, which makes this dataset ideal for evaluating next-generation commonsense reasoning systems.

BibTeX entry

@InProceedings{Bhargava+Ng:22b,
  author = {Prajjwal Bhargava and Vincent Ng},
  title = {{DiscoSense}: Commonsense Reasoning with Discourse Connectives},
  booktitle = {Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing},
  {10295--10310},
  year = 2022}