UMD team wins EMNLP 2026 Best Paper for comparing human and LLM answering strategies
boydgraber · x · 2026-11-10
- UMD CS professors Jordan Boyd-Graber and Rachel Rudinger with PhD student Neha Srikanth won a Best Paper Award at EMNLP 2026 (Budapest, Oct 24–29) — one of only two awarded — for "DiscoTrace: Representing and Comparing Answering Strategies of Humans and LLMs in Information-Seeking Question Answering."
- The work goes beyond factual correctness, analyzing how respondents approach questions: addressing one interpretation vs. covering multiple possibilities, challenging assumptions, or asking for clarification.
- Boyd-Graber: "AI answers are like informational junk food — they sound good, but don't do what they need to do," explaining why AI answers sound similar regardless of question type. The project grew out of earlier research on vague and assumption-laden questions.
More from Research
- RIDArena: A New Benchmark Setting for Evaluating Rogue Internal AI Deployments — jammastergirish · 2026-11-10
- Open-source TotalSegmentator traces airways in 38s on one GPU, navigates to lung nodule — MaziyarPanahi · 2026-11-10
- EEG vs implanted BCIs: researchers clash over physical limits and AI decoding potential — burny_tech · 2026-11-10
- Charlotte Stix traces the history of "trustworthy AI" as a policy concept — ArtificialOther · 2026-11-10
- The Softmax Wall: in FP8/FP4 attention is no longer bottlenecked by matmuls — burny_tech · 2026-11-10
- ExploitGym wins Best Benchmark at Open Source AI Awards, flags OpenAI–Hugging Face incident — dawnsongtweets · 2026-11-10