What do we expect from Multiple-choice QA Systems?
Shah, Krunal, Gupta, Nitish, Roth, Dan
–arXiv.org Artificial Intelligence
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.
arXiv.org Artificial Intelligence
Nov-20-2020
- Country:
- Oceania > Australia
- North America > United States
- Pennsylvania (0.04)
- Louisiana > Orleans Parish
- New Orleans (0.05)
- Europe
- Denmark > Capital Region
- Copenhagen (0.05)
- Belgium > Brussels-Capital Region
- Brussels (0.05)
- Denmark > Capital Region
- Asia > China
- Hong Kong (0.04)
- Genre:
- Research Report > New Finding (0.69)
- Industry:
- Technology: