References & Citations
Computer Science > Computation and Language
Title: CONTaiNER: Few-Shot Named Entity Recognition via Contrastive Learning
(Submitted on 15 Sep 2021 (v1), last revised 28 Mar 2022 (this version, v2))
Abstract: Named Entity Recognition (NER) in Few-Shot setting is imperative for entity tagging in low resource domains. Existing approaches only learn class-specific semantic features and intermediate representations from source domains. This affects generalizability to unseen target domains, resulting in suboptimal performances. To this end, we present CONTaiNER, a novel contrastive learning technique that optimizes the inter-token distribution distance for Few-Shot NER. Instead of optimizing class-specific attributes, CONTaiNER optimizes a generalized objective of differentiating between token categories based on their Gaussian-distributed embeddings. This effectively alleviates overfitting issues originating from training domains. Our experiments in several traditional test domains (OntoNotes, CoNLL'03, WNUT '17, GUM) and a new large scale Few-Shot NER dataset (Few-NERD) demonstrate that on average, CONTaiNER outperforms previous methods by 3%-13% absolute F1 points while showing consistent performance trends, even in challenging scenarios where previous approaches could not achieve appreciable performance.
Submission history
From: Sarkar Snigdha Sarathi Das [view email][v1] Wed, 15 Sep 2021 21:41:16 GMT (245kb,D)
[v2] Mon, 28 Mar 2022 06:47:40 GMT (603kb,D)
Link back to: arXiv, form interface, contact.