Overview
Project Team
Publications
Software

Datasets
Media

CAREER: Human-Computer Cooperation for Word-by-Word Question Answering

Project funded by the National Science Foundation (Originally IIS-1652666 at U Colorado, now 1822494 at UMD)
PI: Jordan Boyd-Graber, University of Maryland

Overview

This CAREER project investigates how humans and computers can work together to answer questions. Humans and computers possess complementary skills: humans have extensive commonsense understanding of the world and greater facility with unconventional language, while computers can effortlessly memorize countless facts and retrieve them in an instant. This proposal helps machines understand who people, places, and characters are; how to communicate this information to humans; and how to allow humans and computers to collaborate in question answering using limited information. A key component of this proposal is answering questions word-by-word: this forces both humans and computers to answer questions using information as efficiently as possible. In addition to embedding these skills in question answering tasks, this proposal has an extensive outreach program to exhibit this technology in interactive question answering competitions for high school and college students.

This research is possible by a new representations of entities in a medium-dimensional embedding that encodes relationships between entities (e.g., the representation of "Goodluck Jonathan" and "Nigeria" encodes that the former is the leader of the latter) to enable the system to answer questions about Nigeria. We validate the effectiveness of these representations both through traditional question answering evaluations and through interactive experiments with human collaboration to ensure that we can visualize these representations effectively. In addition to helping train computers to answer questions, we use opponent modeling and reinforcement learning to help train humans to better answer questions.

<< back to top

Project Team

Jordan Boyd-Graber Jordan Boyd-Graber
Assistant Professor, Computer Science (Maryland)
Ahmed Elgohary
Ph.D. student, Computer Science (Maryland)
Shi Feng Shi Feng
Ph.D. student, Computer Science (Maryland)
Pedro Rodriguez Pedro Rodriguez
Ph.D. student, Computer Science (Maryland)
Eric Wallace Eric Wallace
Undergrad, Computer Science (Maryland)
Chen Zhao Chen Zhao
Ph.D. student, Computer Science (Maryland)

<< back to top

Publications (Selected)

Software

Resources

Media

  • Matt Gardner and Waleed Ammar. A Discussion of Question Answering. AI2 NLP Highlights, 2018. [Bibtex]
  • Matt Gardner and Waleed Ammar. Pathologies of Neural Models Make Interpretation Difficult. AI2 NLP Highlights, 2018. [Bibtex]
  • Matt Early Wright. Inside AI's "Black Box". Maryland Today, 2018. [Bibtex]
  • Sala Levin. What Is ... a Research-Inspired Route to "Jeopardy!"?. Maryland Today, 2018. [Bibtex]
  • Samir Singh. Summary of Pathologies of Neural Models Make Interpretations Difficult. UCI NLP, 2018. [Bibtex]
  • Brandi Adams. Associate Professor Jordan Boyd-Graber to appear on Jeopardy on September 26th 2018. UMD Computer Science, 2018. [Bibtex]
  • Melissa Brachfeld. Boyd-Graber Publishes Paper in PNAS that Assesses Scholarly Influence. UMIACS, 2018. [Bibtex]
  • Workshops Organized by Project Members

    Acknowledgments

    This work is supported by the National Science Foundation. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the researchers and do not necessarily reflect the views of the National Science Foundation.