MINT: Minimal Information Neuro-Symbolic Tree for Objective-Driven Knowledge-Gap Reasoning and Active Elicitation
Zeyu Fang, Tian Lan, Mahdi Imani · 2026
Read MINT: Minimal Information Neuro-Symbolic Tree for Objective-Driven Knowledge-Gap Reasoning and Active Elicitation by Zeyu Fang, Tian Lan, Mahdi Imani (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-05048.txt.
Description
Planning with a human partner means not knowing everything: objects, goals, and intentions can all be missing. Zeyu Fang and 2 co-authors built a system for asking the right questions in that situation in this 2026 artificial intelligence paper. An agent that asks too much wastes time; one that asks too little plans blind.
Their method, MINT, grows a symbolic tree of possible human-AI interactions and uses a neural planning policy to estimate how much each knowledge gap clouds the outcome. Self-play then tunes the agent's questioning strategy, and a language model searches the tree's reasoning to draft the best queries. The authors prove a return guarantee for the approach. On three benchmarks with unknown objects of rising realism, MINT-based planning reached near-expert results while asking only a few questions per task.
Catalog record
| CREATOR | Zeyu Fang, Tian Lan, Mahdi Imani |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-05048