MirrorWiC: On Eliciting Word-in-Context Representations from Pretrained Language Models
This addresses the challenge of enhancing context-aware word understanding in NLP for applications like semantic analysis, though it is incremental as it builds on existing contrastive learning techniques.
The authors tackled the problem of improving word-in-context representations in pretrained language models without supervision, achieving substantial gains across monolingual, multilingual, and cross-lingual benchmarks, with MirrorWiC matching supervised models on some tasks.
Recent work indicated that pretrained language models (PLMs) such as BERT and RoBERTa can be transformed into effective sentence and word encoders even via simple self-supervised techniques. Inspired by this line of work, in this paper we propose a fully unsupervised approach to improving word-in-context (WiC) representations in PLMs, achieved via a simple and efficient WiC-targeted fine-tuning procedure: MirrorWiC. The proposed method leverages only raw texts sampled from Wikipedia, assuming no sense-annotated data, and learns context-aware word representations within a standard contrastive learning setup. We experiment with a series of standard and comprehensive WiC benchmarks across multiple languages. Our proposed fully unsupervised MirrorWiC models obtain substantial gains over off-the-shelf PLMs across all monolingual, multilingual and cross-lingual setups. Moreover, on some standard WiC benchmarks, MirrorWiC is even on-par with supervised models fine-tuned with in-task data and sense labels.