Eric Nyberg

Professor and Director, Master of Computational Data Science Program

Carnegie Mellon University
School of Computer Science
Language Technologies Institute

[ Contact ]

Noted for his contributions to the fields of automatic text translation, information retrieval, and automatic question answering, Nyberg received his Ph.D. from Carnegie Mellon University (1992), and his B.A. from Boston University (1983). In 2012, Nyberg received the Allen Newell Award for Research Excellence for his scientific contributions to the field of question answering and his work as an original developer on the Watson project. He received the BU Computer Science Distinguished Alumna/Alumnus Award on September 27, 2013. Nyberg currently serves as Director of the Master of Computational Data Science (MCDS) degree program, and is co-Founder and Chief Data Scientist at Cognistx.

Publications and Patents

Eric Nyberg's Google Scholar profile ( most cited | most recent )

Current Courses

11-791 : Intelligent Information Systems (Fall, Spring), 12 units
11-792 : Intelligent Information Systems Project (Fall, Spring), 12 units
11-796 : Question Answering Lab (Spring), 6 units
11-797 : Question Answering (Spring), 12 units

Research: Open Advancement of Question Answering

The Open Advancement of Question Answering (OAQA) project (2007-present) works to provide foundational architecture and methodology for agile collaborative research in automatic question answering. By making long-term commitments to an architecture that supports best-practices research methodology, OAQA is accelerating the advancement of QA technologies in several application areas. The OAQA approach combines an object-oriented software architecture with comprehensive metrics, measurement and error analysis, at the system and module levels, and has been applied to a variety of challenge problems (see below for notable examples).

The LiveQA Challenge. Beginning in 2015, CMU has collaborated with Yahoo! Labs (as part of the InMind project) to develop automatic answering agents that can respond to real-time questions from web users (like those received by the Yahoo! Answers community QA web site). CMU student Di Wang created a LiveQA system which combined standard retrieval algorithms (BM25) with state-of-the-art deep learning models [1] to achieve the highest score among all participants in the 2015 TREC LiveQA Challenge [2]. In 2016, Di extended his system to include a novel answer ranking method based on attentional encoder-decoder recurrent neural networks [3] and achieved the highest score among 25 automatic systems that were evaluated in the 2016 LiveQA Track [4].

[1] D. Wang and E. Nyberg (2015). "CMU OAQA at TREC 2015 LiveQA: Discovering the Right Answer with Clues", Proceedings of TREC 2015 [ PDF ]

[2] E. Agichtein, D. Carmel, D. Harman, D. Pelleg, Y. Pinter (2015). "Overview of the TREC 2015 Live QA Track", Proceedings of TREC 2015 [ PDF ]

[3] D. Wang and E. Nyberg (2016). "CMU OAQA at TREC 2016 LiveQA: An Attentional Neural Encoder-Decoder Approach for Answer Ranking", Proceedings of TREC 2016 [ PDF ]

[4] E. Agichtein, D. Carmel, D. Pelleg, Y. Pinter and D. Harman (2016). "Overview of the TREC 2016 Live QA Track", Proceedings of TREC 2016

The BioASQ Challenge. From 2012 to 2016, a team led by LTI Ph.D. student Zi Yang collaborated with Hoffman-LaRoche's Innovation Center to develop information systems for unstructured biomedical text, including a passage retrieval system for the TREC Genomics dataset [1], a decision support system for gene targeting which leverages information gathered from PubMed articles by an automatic QA system [2], a Biomedical Semantic QA system which received six 1st-place scores in the 2015 BioASQ Challenge tasks, which included snippet retrieval, concept retrieval, and exact answer retrieval [3], and a Biomedical Semantic QA system which received three 1st-place scores in exact answer retrieval in the 2016 BioASQ Challenge [4].

[1] Z. Yang, E. Garduno, Y. Fang, A. Maiberg, C. McCormack, and E. Nyberg (2013). "Building Optimal Information Systems Automatically: Configuration Space Exploration for Biomedical Information Systems", Proceedings of the ACM Conference on Information and Knowledge Management [ ACM Digital Library ]

[2] Z. Yang, Y. Li, J. Cai, and E. Nyberg (2014). "QUADS: Question Answering for Decision Support." In Proceedings of SIGIR 2014: the Thirty-seventh Annual International ACM SIGIR Conference on Research and Development in Information Retrieval [ ACM Digital Library ]

[3] Z. Yang, N. Gupta, X. Sun, D. Xu, C. Zhang, and E. Nyberg (2015). "Learning to Answer Biomedical Factoid & List Questions: OAQA at BioASQ 3B", In Proceedings of CLEF 2015 Evaluation Labs and Workshop [ PDF ]

[4] Z. Yang, Y. Zhou and E. Nyberg (2016). "Learning to Answer Biomedical Questions: OAQA at BioASQ 4B", In Proceedings of Workshop on Biomedical Language Processing [ PDF ]

The Jeopardy! Challenge. From 2007 to 2011, CMU collaborated with the IBM DeepQA Group to develop an open-source framework for OAQA [1]. The initial OAQA architecture and data model were used to build systems for the TREC challenge problem and the Jeopardy! challenge problem [2,3,4]. Carnegie Mellon students Nico Schlaefer and Hideki Shima also contributed algorithms and code to IBM's Watson system, as participants in IBM's summer internship program.

[1] "Towards the Open Advancement of Question Answering", IBM Technical Report, 2008

[2] "CMU and IBM Collaborate on Open Computing System for Question Answering", PR Newswire, 02/11/11

[3] "IBM Announces Eight Universities Contributing to Watson", PR Newswire, 02/11/11

[4] "Man versus machine: Chalk one up for the latter in Jeopardy! showdown'", Pittsburgh Post-Gazette, 02/17/11

Last Updated 17-June-2017