HuggingFace Daily Papers·· 1 天前
SHELDON：在假设空间与观测空间上双重搜索的推理框架
SHELDON: Reasoning through Dual Search over Hypothesis and Observation Spaces
AI 导读
SHELDON 是一种受假说-演绎法与科学发现双空间模型启发的科学推理框架，将假设空间与观测空间分开表示并联合探索，从而发现当前假设无法解释的观测。在诊断、抽象推理与优化基准上，它最多超过基线 26.7 分，并生成最多 4.5 倍的新颖解。
正文
Abstract:Test-time scaling has enabled LLMs to solve increasingly complex reasoning, discovery, and optimization tasks. However, existing approaches often gather observations while seeking evidence for current hypotheses, then generate new hypotheses conditioned on those hypotheses and their associated evidence. This couples observation discovery to existing hypotheses, favoring local refinements over exploration of unknown hypotheses regions. Inspired by the hypothetico-deductive method (Hempel, 1945) and dual-space models of scientific discovery (Klahr & Dunbar, 1988), we introduce SHELDON, a scientific reasoning harness that separately represents and jointly explores hypothesis and observation spaces. By actively seeking new observations along with new hypotheses, SHELDON can discover observations that its current hypotheses do not explain. These unexplained observations expose gaps in its understanding, motivate new investigations, and guide the generation of hypotheses beyond refinements of existing ones. Across diagnosis, abstract reasoning, and optimization benchmarks, SHELDON outperforms evaluated baselines by up to 26.7 points. Additional analysis shows that SHELDON generates up to 4.5x more novel solutions.
| Comments: | 16 pages main text, 34 total. Preprint |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.18025 [cs.AI] |
| (or arXiv:2610.18025v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.18025 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Harshit Joshi [view email]
[v1]
Tue, 13 Oct 2026 09:19:23 UTC (4,036 KB)
来源：HuggingFace Daily Papers · arxiv.org