Gaze as evidence for common grounding: cross-corpus analysis of MapTask and MUNDEX
cs-nlp-uu · hf · 2026-09-17
This study asks whether gaze provides evidence of grounding in asymmetric-information collaborative tasks. Mapping HCRC MapTask and MUNDEX into a shared partner/task/away vocabulary, the authors find that aligned references and 'understood' judgments are associated with more task-directed gaze, less partner-directed gaze, lower gaze entropy, and fewer transitions—strongest for the task-leading participant. Effects are small and weaken when recurring participants are the unit of inference, so gaze should be treated as one contributing cue alongside task and dialogue context.
More from Research
- ScienceIDE Turns the World's Scientific Codebases into Agent-Learnable Environments — Hejia Geng · 2026-09-17
- EventEgoHands++: Event-Camera Egocentric 3D Hand Reconstruction with 1M-Frame Real Dataset — Ryosei Hara · 2026-09-17
- EvolveTrade: Self-Evolving LLM Trading Agents Refine Their Own Tool-Use Policy — kaist-ai · 2026-09-17
- Tobias Lee: RL for Verifiable Tasks, MOPD for Open Domains — _AndrewZhao · 2026-09-17
- Schmidhuber Publishes 40-Year Retrospective on Recursive Self-Improvement Since 1987 — rbhar90 · 2026-09-17
- LocalFold runs AlphaFold3, Boltz-2 and more protein folding models in your browser via WebGPU — sokrypton · 2026-09-17