Numerically Grounded Language Models for Semantic Error Correction

TL;DR

This paper introduces numerically grounded language models that improve semantic error detection and correction, especially in clinical texts, by incorporating numeric information and background knowledge bases, leading to significant performance gains.

Contribution

It presents a novel approach that grounds language models in numeric data and background knowledge, enhancing semantic error correction capabilities.

Findings

Numerical grounding improves perplexity by 33%.

Semantic error correction F1 score increases by 5 points.

Conditioning on knowledge bases yields further improvements.

Abstract

Semantic error detection and correction is an important task for applications such as fact checking, speech-to-text or grammatical error correction. Current approaches generally focus on relatively shallow semantics and do not account for numeric quantities. Our approach uses language models grounded in numbers within the text. Such groundings are easily achieved for recurrent neural language model architectures, which can be further conditioned on incomplete background knowledge bases. Our evaluation on clinical reports shows that numerical grounding improves perplexity by 33% and F1 for semantic error correction by 5 points when compared to ungrounded approaches. Conditioning on a knowledge base yields further improvements.