AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Empirical Analysis of Unlabeled Entity Problem in Named Entity Recognition

Yangming Li, lemao liu, Shuming Shi

ICLR 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 7 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use negative sampling instead of treating all unlabeled spans as negativesBERT-based span NER model · CoNLL-2003 synthetic, masking 0.6F1: 63.84 -> 87.65Improved
use negative sampling instead of treating all unlabeled spans as negativesBERT-based span NER model · OntoNotes 5.0 synthetic, masking 0.6F1: 58.17 -> 87.53Improved
use negative sampling instead of treating all unlabeled spans as negativesBERT-based span NER model · ECF1: 63.56 -> 66.17Improved
use negative sampling instead of treating all unlabeled spans as negativesBERT-based span NER model · NEWSF1: 80.04 -> 85.39Improved
use negative sampling on fully annotated dataBERT-based span NER model · CoNLL-2003F1: 93.5 -> 93.42Hurt
use negative sampling on fully annotated dataBERT-based span NER model · OntoNotes 5.0F1: 91.3 -> 90.59Hurt
replace the BERT encoder with BiLSTM (ablation)Our Model · ECF1: 66.17 -> 64.68Hurt
replace the BERT encoder with BiLSTM (ablation)Our Model · NEWSF1: 85.39 -> 82.11Hurt
vary the negative-sampling ratio λ (ablation)Our Model · synthetic datasetsF1Mixed
use a pretrained language model rather than LSTM to mitigate fewer annotated entitiestagging model with adjusted loss · CoNLL-2003 and OntoNotes 5.0 synthetic, masking 0.8erosion rate: about halfImproved
treat unlabeled entities as training negativesNER models · synthetic datasetsF1: about 20%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy