Chin-Yew Lin

Also known as Chinyew Lin

Chin-Yew Lin is a natural language processing researcher at Microsoft Research Asia who created the ROUGE automatic summarization evaluation package. After earning a Ph.D. at the University of Southern California in 1997, he worked as a senior research scientist in the Natural Language Processing and Machine Translation group at USC's Information Sciences Institute and published a topic-identification paper with Eduard Hovy in 1997. With Hovy, he published a summary-evaluation paper using n-gram co-occurrence statistics in 2003, introduced ROUGE in July 2004, and presented the Basic Elements framework at LREC in 2006. He served as program co-chair of ACL 2012, led the Microsoft team that shipped the Entity Linking Intelligence Service in 2016, coauthored "List-only Entity Linking" in 2017 and the LLMLingua and LongLLMLingua prompt-compression papers in 2023 and 2024, and is a Senior Principal Research Manager leading the Knowledge Computing group at Microsoft Research Asia.

Now at: Microsoft Research Asia
Badges: Bridge

Timeline

1997

A USC doctorate and ISI post

He earns a Ph.D. at the University of Southern California in 1997 and works from 1997 as a senior research scientist in the Natural Language Processing and Machine Translation group at USC's Information Sciences Institute before joining Microsoft. At the Fifth Conference on Applied Natural Language Processing in Washington, DC, he and Eduard Hovy publish "Identifying Topics by Position" in 1997.
2003–2004

Creating the ROUGE package

With Eduard Hovy, he publishes "Automatic Evaluation of Summaries Using N-gram Co-occurrence Statistics" in 2003, and in July 2004 he presents "ROUGE: A Package for Automatic Evaluation of Summaries" at the Workshop on Text Summarization Branches Out in Barcelona, Spain. From 2004, the Document Understanding Conference uses the ROUGE automatic summarization evaluation package as its official automatic evaluation package. A 2004 ACL paper with Franz Josef Och applies longest common subsequence and skip-bigram statistics to automatic evaluation of machine translation quality.
2006

Basic Elements for summary scoring

Together with Eduard Hovy, Liang Zhou and Junichi Fukumoto, he presents "Automated Summarization Evaluation with Basic Elements" at the Fifth International Conference on Language Resources and Evaluation (LREC) in Genoa, Italy, in 2006.
2011–2016

Program co-chair of three conferences

He serves as program co-chair of the AAAI 2011 AI & the Web Special Track, of ACL 2012 in Jeju, Republic of Korea, with Miles Osborne, and of NLPCC 2016.
2013–2016

Entity linking at Microsoft

His team at Microsoft scores the best F1 in the Knowledge Base Acceleration Evaluation in 2013 and 2014 and ships the Entity Linking Intelligence Service (ELIS) at Microsoft //BUILD in 2016.
2017

List-only entity linking at ACL

With Ying Lin and Heng Ji, he publishes "List-only Entity Linking" at the 55th Annual Meeting of the Association for Computational Linguistics in 2017.
2023–2024

Prompt compression with LLMLingua

In 2023, he coauthors "LLMLingua: Compressing Prompts for Accelerated Inference of Large Language Models" with Huiqiang Jiang, Qianhui Wu, Yuqing Yang and Lili Qiu, introducing a prompt compression approach that allows up to 20x compression with little performance loss. In 2024, he coauthors LongLLMLingua, which boosts performance on the NaturalQuestions benchmark by up to 21.4% with around 4x fewer tokens in GPT-3.5-Turbo.
As of October 2026

Senior Principal Research Manager

He is a Senior Principal Research Manager at Microsoft Research Asia and a research manager of its Knowledge Computing group. His research interests cover knowledge computing, natural language processing, semantic search, text generation, question answering, and automatic summarization.