AI Paper Atlas

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

Sequence Level Contrastive Learning for Text Summarization

Shusheng Xu, Xingxing Zhang, Yi Wu, Furu Wei

AAAI 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 78 citations (OpenAlex)

This paper tried a common change: Add a contrastive loss

Add contrastive learning · CNNDM and NYT human evaluation · Mean faithfulness rank: 2.62 -> 2.42; 2.68 -> 2.39 Improved

What 1,317 accepted papers report for this change

See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.

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
Add sequence-level contrast between document and gold summaryBART Large on XSum · XSum testROUGE-1: 45.35 -> 45.65Improved
Add sequence-level contrast between document and gold summaryBART Large on XSum · XSum testROUGE-2: 22.01 -> 22.41Improved
Add sequence-level contrast between document and gold summaryBART Large on XSum · XSum testROUGE-L: 36.76 -> 37.04Improved
Add contrast between document and generated summaryBART Large on NYT · NYT testROUGE-1: 53.20 -> 54.25Improved
Add contrast between document and generated summaryBART Large on NYT · NYT testROUGE-2: 35.04 -> 35.82Improved
Add contrast between document and generated summaryBART Large on NYT · NYT testROUGE-L: 49.23 -> 50.24Improved
Use generated summaries rather than gold summaries as contrast partnersSeqCo on CNNDM and NYT · CNNDM and NYT testROUGEImproved
Use generated summaries rather than gold summaries as contrast partnersSeqCo on XSum · XSum testROUGEHurt
Replace multi-head attention similarity with [CLS] pooling (ablation)SeqCo on CNNDM · CNNDM validationROUGE-1: 45.88 -> 45.72Hurt
Add multiple contrastive pairs instead of one (ablation)SeqCo on CNNDM · CNNDM validationROUGE-1: 45.88 -> 45.72Hurt
Use decoder-conditioned representations for contrastive learning (ablation)SeqCo on CNNDM · CNNDM validationROUGE-L: 42.66 -> 41.55Hurt
Add multiple contrastive pairs instead of one (ablation)SeqCo on XSum · XSum validationROUGE-2: 22.42 -> 22.52Improved

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