BEGIN:VCALENDAR VERSION:2.0 PRODID:-//wp-events-plugin.com//7.4.5//EN TZID:America/Los_Angeles X-WR-TIMEZONE:America/Los_Angeles BEGIN:VEVENT UID:378@escience.washington.edu DTSTART;TZID=America/Los_Angeles:20261013T163000 DTEND;TZID=America/Los_Angeles:20261013T172000 DTSTAMP:20261008T213225Z URL:https://escience.washington.edu/events/uw-data-science-seminar-justin- payan/ SUMMARY:UW Data Science Seminar: Justin Payan DESCRIPTION:\nPlease join us for a UW Data Science Seminar featuring Carneg ie Mellon Machine Learning Postdoctoral Scholar Justin Payan on Tuesday\, October 13th from 4:30 to 5:20 p.m. PT. The seminar will be held in HCK 13 2.\n“AI in Science: Building Infrastructure for an Explosion of Discover y”\nAbstract: AI is rapidly accelerating math and science: just last mon th an internal OpenAI model solved the Navier-Stokes Millennium Prize prob lem\, and Anthropic’s Claude drove the discovery of a promising enzyme s ystem in a wet lab. Yet AI also allows anyone with an internet connection to generate ostensibly correct pseudoscience at the click of a button. Our scientific infrastructure is buckling under the increased volume.\nI prop ose leveraging AI at key points to focus human oversight\, and I demonstra te two applications of AI in peer review. Many peer review venues have dep loyed prototype AI reviewers to serve alongside human experts\, claiming t hat AI reviewers can help identify errors and weed out obviously poor qual ity submissions. But do these AI reviewers accurately identify problems in submissions? I present a benchmark called FLAWS (Fault Localization Acros s Writing in Science)\, which inserts major flaws into papers to evaluate AI reviewer capabilities. At the time of evaluation (November 2025)\, we f ound that the best-performing model (GPT-5) identified inserted errors in <\;40% of instances. The journal Transactions on Machine Learning Resear ch (TMLR) also used FLAWS to evaluate AI reviewers\, finding the top-perfo rming system’s performance satisfactory for experimental deployment alon gside human reviewers.\nEven with AI reviewers\, peer review venues still struggle to keep up with increasing submissions of AI-generated articles – articles that the “authors” may not have even read! We develop a t est called greCAPTCHA\, which verifies author understanding by asking auth ors questions about their own papers. In a study with 31 researchers\, we show that we are able to distinguish researchers answering questions about their own versus unfamiliar papers with an ROC AUC of 0.9. Semi-structure d interviews show broad support for greCAPTCHA\, and suggest changes to ma ke before deployment. A public demo is available at grecaptcha.com. These projects examine how AI can support scientific integrity\, while preservin g human judgment and accountability.\nSpeaker Bio: Justin Payan is a Postd octoral Research Associate in the Machine Learning Department at Carnegie Mellon University. He applies machine learning\, combinatorial optimizatio n\, and the social sciences to improve scientific integrity and efficiency . He has published papers in IJCAI\, AAMAS\, NeurIPS\, WSDM\, and EMNLP. H is reviewer assignment algorithm\, FairSequence\, has been used by over 50 venues on OpenReview to achieve fast and fair reviewer assignments. His w ork assessing author understanding was recently covered by Times Higher Ed ucation\, and his collaboration with the ALMA observatory in Chile led to a novel telescope scheduling algorithm currently being evaluated for deplo yment.\n\n\nThe 2026-2027 seminars will be held in person\, and are free a nd open to the public.\n\n\n\n\n\n\n ATTACH;FMTTYPE=image/jpeg:https://escience.washington.edu/wp-content/uploa ds/2026/10/image.jpeg END:VEVENT BEGIN:VTIMEZONE TZID:America/Los_Angeles X-LIC-LOCATION:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 TZNAME:PDT END:DAYLIGHT END:VTIMEZONE END:VCALENDAR