 |
Director: Prof. Matthew Lease Publications • Datasets & Software • Slides 1616 Guadalupe Ste 5.202 Lab Photo |
Research Areas (specific): Information Retrieval (IR) •
Crowdsourcing & Human Computation (HCOMP) •
Natural Language Processing (NLP)
Research Areas (general): Artificial Intelligence (AI) • Human-Computer Interation (HCI)
About the Lab: Overview • 5-slides
Overview: IR is the science behind search engines such as Google, Bing, & Yahoo. Crowdsourcing and human computation engage online workers to train or augment automated artificial intelligence algorithms. My IR research seeks to improve core search algorithms, reliably evaluate search systems, and to enable new forms of search. My HCOMP research seeks to optimize crowdsourced data collection (e.g., quality, cost, and speed), to expand the reach of crowdsourcing to tackle new problems, and to investigate broader socio-technical questions of how paid crowdsourcing is transforming digital work and the lives of workers engaged in it. At the intersection of IR and HComp, I develop crowdsourcing methods to better scale IR evaluation methodology while preserving its reliability. Both IR and HComp place people at the center of computing: system users in IR and online workers in HComp. I thus seek to orchestrate effective man-machine partnerships which creatively blend front-end HCI design with back-end AI modeling of people and their tasks. By capitalizing on the respective strengths of each party - man and machine - we can compensate for the other's limitations to create a whole greater than the sum of its parts. For example, IR systems can utilize front-end HCI design to empower searcher intuition and creativity, while back-end AI algorithms interpret ambiguous human queries, sift through vast information, and suggest potentially relevant results. In HCOMP, front-end HCI design can enable workers to more easily understand and complete tasks, while back-end AI modeling of workers and tasks enables principled optimization of data collection.
Selected Research & Demos
Anubrata Das, Brandon Dang, and Matthew Lease. Fast, Accurate, and Healthier: Interactive Blurring Helps Moderators Reduce Exposure to Harmful Content. In Proceedings of the 8th AAAI Conference on Human Computation and Crowdsourcing (HCOMP), pages 33--42, 2020. [ bib | pdf | demo | blog-post | sourcecode | video | slides ]
Mucahid Kutlu, Tyler McDonnell, Tamer Elsayed, and Matthew Lease. Annotator Rationales for Labeling Tasks in Crowdsourcing. Journal of Artificial Intelligence Research (JAIR), 69:143--189, 2020. Award Winning Papers Track. [ bib | pdf | blog-post | data | conference-website ]
Soumyajit Gupta, Mucahid Kutlu, Vivek Khetan, and Matthew Lease. Correlation, Prediction and Ranking of Evaluation Metrics in Information Retrieval. In Proceedings of the 41st European Conference on Information Retrieval (ECIR), pages 636--651, 2019. Best Student Paper award. [ news | bib | pdf | data | sourcecode | slides | tech-report ]
An Thanh Nguyen, Aditya Kharosekar, Aditya Kharosekar, Saumyaa Krishnan,
Siddhesh Krishnan, Elizabeth Tate, Byron C. Wallace, and Matthew Lease.
Believe it or not: Designing a Human-AI Partnership for
Mixed-Initiative Fact-Checking.
In Proceedings of the 31st ACM User Interface Software and
Technology Symposium (UIST), pages 189--199, 2018.
[ bib |
pdf |
demo |
sourcecode |
video |
slides ]
Tyler McDonnell, Matthew Lease, Mucahid Kutlu, and Tamer Elsayed. Why Is That Relevant? Collecting Annotator Rationales for Relevance Judgments. In Proceedings of the 4th AAAI Conference on Human Computation and Crowdsourcing (HCOMP), pages 139--148, 2016. Best Paper Award. [ news | bib | pdf | blog-post | data | slides ]
Hyun Joon Jung and Matthew Lease. A Discriminative Approach to Predicting Assessor Accuracy. In Proceedings of the 37th European Conference on Information Retrieval (ECIR), pages 159--171, 2015. Samsung Human-Tech Paper Award: Silver Prize in Computer Science. [ bib | pdf | news ]
Aashish Sheshadri and Matthew Lease. SQUARE: A Benchmark for Research on Computing Crowd Consensus. In Proceedings of the 1st AAAI Conference on Human Computation (HCOMP), pages 156--164, 2013. [ bib | pdf | data ]
Talks and Slides
News
Current PhD Studets (as of Fall 2020)
Alumni
Join our Lab!
What's it like in an Information School (iSchool)? See Wobbrock et al.'s short manifesto
UT Austin iSchool Facts
Prospective graduate students: Graduate Study in IR at UT Austin
Current MSIS students
High School Students: apply for UT's Summer Research Academy
