Infusing Finetuning with Semantic Dependencies

TL;DR

This paper investigates the limitations of current pretrained language models in capturing semantic dependencies and introduces a method to explicitly incorporate semantic parses during finetuning, improving performance on NLU tasks.

Contribution

The paper presents a novel approach using convolutional graph encoders to embed semantic parses into finetuning, enhancing language understanding beyond traditional pretraining.

Findings

Semantic dependencies are not well captured by current models

Explicit semantic supervision improves NLU task performance

Diagnostics identify where semantic benefits are most significant

Abstract

For natural language processing systems, two kinds of evidence support the use of text representations from neural language models "pretrained" on large unannotated corpora: performance on application-inspired benchmarks (Peters et al., 2018, inter alia), and the emergence of syntactic abstractions in those representations (Tenney et al., 2019, inter alia). On the other hand, the lack of grounded supervision calls into question how well these representations can ever capture meaning (Bender and Koller, 2020). We apply novel probes to recent language models -- specifically focusing on predicate-argument structure as operationalized by semantic dependencies (Ivanova et al., 2012) -- and find that, unlike syntax, semantics is not brought to the surface by today's pretrained models. We then use convolutional graph encoders to explicitly incorporate semantic parses into task-specific finetuning, yielding benefits to natural language understanding (NLU) tasks in the GLUE benchmark. This approach demonstrates the potential for general-purpose (rather than task-specific) linguistic supervision, above and beyond conventional pretraining and finetuning. Several diagnostics help to localize the benefits of our approach.