Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.19792 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914490719666176 |
|---|---|
| author | Zeng, Peter Li, Weiling Paige, Amie Wang, Zhengxiang Kaliosis, Panagiotis Samaras, Dimitris Zelinsky, Gregory Brennan, Susan Rambow, Owen |
| author_facet | Zeng, Peter Li, Weiling Paige, Amie Wang, Zhengxiang Kaliosis, Panagiotis Samaras, Dimitris Zelinsky, Gregory Brennan, Susan Rambow, Owen |
| contents | For generative AI agents to partner effectively with human users, the ability to accurately predict human intent is critical. But this ability to collaborate remains limited by a critical deficit: an inability to model common ground. We present a referential communication experiment with a factorial design involving director-matcher pairs (human-human, human-AI, AI-human, and AI-AI) that interact with multiple turns in repeated rounds to match pictures of objects not associated with any obvious lexicalized labels. We show that LVLMs cannot interactively generate and resolve referring expressions in a way that enables smooth communication, a crucial skill that underlies human language use. We release our corpus of 356 dialogues (89 pairs over 4 rounds each) along with the online pipeline for data collection and the tools for analyzing accuracy, efficiency, and lexical overlap. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_19792 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | LVLMs and Humans Ground Differently in Referential Communication Zeng, Peter Li, Weiling Paige, Amie Wang, Zhengxiang Kaliosis, Panagiotis Samaras, Dimitris Zelinsky, Gregory Brennan, Susan Rambow, Owen Computation and Language Artificial Intelligence Human-Computer Interaction For generative AI agents to partner effectively with human users, the ability to accurately predict human intent is critical. But this ability to collaborate remains limited by a critical deficit: an inability to model common ground. We present a referential communication experiment with a factorial design involving director-matcher pairs (human-human, human-AI, AI-human, and AI-AI) that interact with multiple turns in repeated rounds to match pictures of objects not associated with any obvious lexicalized labels. We show that LVLMs cannot interactively generate and resolve referring expressions in a way that enables smooth communication, a crucial skill that underlies human language use. We release our corpus of 356 dialogues (89 pairs over 4 rounds each) along with the online pipeline for data collection and the tools for analyzing accuracy, efficiency, and lexical overlap. |
| title | LVLMs and Humans Ground Differently in Referential Communication |
| topic | Computation and Language Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2601.19792 |