AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Lexical Simplification with Pretrained Encoders

Jipeng Qiang, Yun Li, Yi Zhu, Yunhao Yuan, Xindong Wu

AAAI 2020 · Architectures and efficiency › Attention and transformers · approach: Language model pretraining and induction · 94 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
Generate candidates using masked BERT with the original sentence as context, rather than word-embedding similarityBERT-LS candidate generation · LexMTurkF1: 0.195 -> 0.259Improved
Generate candidates using masked BERT with the original sentence as context, rather than word-embedding similarityBERT-LS candidate generation · BenchLSF1: 0.236 -> 0.272Improved
Generate candidates using masked BERT with the original sentence as context, rather than word-embedding similarityBERT-LS candidate generation · NNSevalF1: 0.157 -> 0.218Improved
Use BERT-LS instead of the strongest full-pipeline baselineFull lexical simplification pipeline · LexMTurkAccuracy: 0.682 -> 0.776Improved
Use BERT-LS instead of the strongest full-pipeline baselineFull lexical simplification pipeline · BenchLSAccuracy: 0.434 -> 0.607Improved
Use BERT-LS instead of the strongest full-pipeline baselineFull lexical simplification pipeline · NNSevalAccuracy: 0.335 -> 0.423Improved
Use BERT-LS rather than the strongest full-pipeline baselineFull lexical simplification pipeline · BenchLS and NNSevalPrecision: 0.642 -> 0.607; 0.544 -> 0.423Hurt
Increase the number of candidate substitutions (ablation)BERT-LS candidate generation · LexMTurk, BenchLS, and NNSevalPrecisionHurt
Increase the number of candidate substitutions (ablation)BERT-LS candidate generation · LexMTurk, BenchLS, and NNSevalRecallImproved
Vary the number of candidate substitutions from 1 to 15 (ablation)BERT-LS candidate generation · LexMTurk, BenchLS, and NNSevalF1Mixed
Vary the number of candidate substitutions from 1 to 15 (ablation)BERT-LS full pipeline · LexMTurk, BenchLS, and NNSevalAccuracyMixed

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