Semantics Altering Modifications for Evaluating Comprehension in Machine Reading
This work addresses a critical limitation in current MRC models' ability to truly comprehend text, rather than just pattern match, which is important for developing more robust and reliable NLP systems.
This paper investigates whether state-of-the-art machine reading comprehension (MRC) models can correctly process Semantics Altering Modifications (SAMs), which change sentence semantics while preserving lexical form. They found that optimized models consistently struggle with semantically altered data, despite their high performance on standard benchmarks.
Advances in NLP have yielded impressive results for the task of machine reading comprehension (MRC), with approaches having been reported to achieve performance comparable to that of humans. In this paper, we investigate whether state-of-the-art MRC models are able to correctly process Semantics Altering Modifications (SAM): linguistically-motivated phenomena that alter the semantics of a sentence while preserving most of its lexical surface form. We present a method to automatically generate and align challenge sets featuring original and altered examples. We further propose a novel evaluation methodology to correctly assess the capability of MRC systems to process these examples independent of the data they were optimised on, by discounting for effects introduced by domain shift. In a large-scale empirical study, we apply the methodology in order to evaluate extractive MRC models with regard to their capability to correctly process SAM-enriched data. We comprehensively cover 12 different state-of-the-art neural architecture configurations and four training datasets and find that -- despite their well-known remarkable performance -- optimised models consistently struggle to correctly process semantically altered data.