PACE：自适应猜想探索的研究证明智能体
PACE: Research Proving Agent for Adaptive Conjecture Exploration
AI 导读
PACE 是一个研究证明智能体，通过结构化研究循环选择数学目标、构建并质疑候选机制，由独立验证器核查证明，并借助证据追踪记忆在猜想变化时标记需重新验证的前提。作者同时推出应用数学基准 ResearchProofBench，包含 51 个已知证明任务（331 个隐藏检查点）和 6 个开放问题案例。
正文 · 原文
Abstract:Many mathematical research agents focus on solving problems with fixed statements. However, conjecture research, especially in applied mathematics, may require revising candidate questions or conclusions as investigation proceeds, coupling problem formulation with proof development. We present PACE, a research-proof agent that supports this process through a structured research loop that selects mathematical objectives and develops or challenges candidate mechanisms. The whole prover constructs proofs from these mechanisms, and an independent verifier checks them against the original target. Evidence-tracking memory stores assumptions and dependencies arising during research and marks evidence for revalidation when a candidate or condition changes. This memory supports task-specific context composition that keeps checked premises, exploratory material and unresolved obligations distinct. To evaluate proof construction and conjecture research, we introduce ResearchProofBench, an applied-mathematics benchmark focused on optimization, comprising 51 known-proof tasks with 331 hidden checkpoints under historical abstract access and six open-problem cases. The known-proof tasks assess reconstruction of published results, while the open-problem cases assess mathematical investigation and proof development. On the 51 known-proof tasks, PACE achieves an aggregate Proof Score of 75.16 and completes 20 proofs. In conjecture research, PACE develops conditional results, with its submissions receiving higher review scores than the compared baselines.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.33690 [cs.AI] |
| (or arXiv:2610.33690v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.33690 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Shu Miao [view email]
[v1]
Sat, 24 Oct 2026 14:13:06 UTC (1,927 KB)
来源：arXiv cs.AI · arxiv.org