English

What do we expect from Multiple-choice QA Systems?

Computation and Language 2020-11-24 v1 Artificial Intelligence

Abstract

The recent success of machine learning systems on various QA datasets could be interpreted as a significant improvement in models' language understanding abilities. However, using various perturbations, multiple recent works have shown that good performance on a dataset might not indicate performance that correlates well with human's expectations from models that "understand" language. In this work we consider a top performing model on several Multiple Choice Question Answering (MCQA) datasets, and evaluate it against a set of expectations one might have from such a model, using a series of zero-information perturbations of the model's inputs. Our results show that the model clearly falls short of our expectations, and motivates a modified training approach that forces the model to better attend to the inputs. We show that the new training paradigm leads to a model that performs on par with the original model while better satisfying our expectations.

Keywords

Cite

@article{arxiv.2011.10647,
  title  = {What do we expect from Multiple-choice QA Systems?},
  author = {Krunal Shah and Nitish Gupta and Dan Roth},
  journal= {arXiv preprint arXiv:2011.10647},
  year   = {2020}
}

Comments

Findings of EMNLP 2020

R2 v1 2026-06-23T20:24:26.558Z