Search papers, labs, and topics across Lattice.
Affiliation:
3
0
3
2
This work maps HCRC MapTask and MUNDEX into a shared partner/task/away vocabulary and compute gaze features around task-relevant dialogue units and treats gaze as one contributing cue to grounding, to be interpreted alongside task and dialogue context.
LLMs not only understand plural references but also mirror human preferences in pronoun usage based on entity similarity and conjunctions.
Vision-language models may misinterpret potential common ground as actual understanding, leading to significant biases in collaborative dialogue tasks.
