Accurate, yet inconsistent? Consistency Analysis on Language Understanding Models
This addresses the reliability and trustworthiness of language models for users in NLP applications, but it is incremental as it focuses on evaluating and slightly improving an existing property.
The paper tackles the problem of inconsistent predictions in pretrained language models for semantically similar inputs, finding that current models are prone to inconsistency and that multi-task training with paraphrase identification tasks improves consistency by 13% on average.
Consistency, which refers to the capability of generating the same predictions for semantically similar contexts, is a highly desirable property for a sound language understanding model. Although recent pretrained language models (PLMs) deliver outstanding performance in various downstream tasks, they should exhibit consistent behaviour provided the models truly understand language. In this paper, we propose a simple framework named consistency analysis on language understanding models (CALUM)} to evaluate the model's lower-bound consistency ability. Through experiments, we confirmed that current PLMs are prone to generate inconsistent predictions even for semantically identical inputs. We also observed that multi-task training with paraphrase identification tasks is of benefit to improve consistency, increasing the consistency by 13% on average.