arXiv cs.AI· Silv\`ere Gangloff, Jan Hula·· 3 天前AI 评分24
简单离散世界之间的知识迁移
Knowledge transfer between simple discrete worlds
AI 导读
研究在类数独的小型离散世界中测试深度循环模型的知识迁移能力，训练单一模型掌握多种规则组合，并在四个新颖性层级上评估。在重组已训练规则时，符号描述达到99%、示例为95%；当规则组件从未在训练中共现时，描述方式低于无信息基线，而示例方式仍高于基线。该框架比 ARC-AGI 等基准更系统地测试泛化能力，为评估深度循环推理模型提供新依据。
正文 · 原文
Abstract:Adapting to a new world and reasoning correctly about which outcomes are possible under rules one has not met before is a basic capacity underlying general intelligence. We study it in small discrete worlds: sets of grid configurations consistent with a shared set of rules, concretely Sudoku-like puzzles built from a common base and a handful of extra rules. We train a single depth-recurrent model on many combinations of rules and test it at four levels of novelty: a new puzzle, a new combination of known rules, a pairing of rules never seen together, and a rule instance not seen at all during training. The model receives the active rules either as a symbolic description or as two solved examples, and a model that gets no rule information is used as a baseline. At recombining rules it was trained on, the description matches or exceeds examples ($99\%$ against $95\%$ on new combinations). On rule instances whose components never co-occurred in training, the description falls below the no-information baseline while examples stay above it. The framework allows us to test many such generalization regimes in more systematic way than benchmarks such as ARC-AGI and provides new ground for evaluating depth-recurrent reasoning models, which are typically tested on saturated Sudoku benchmarks.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.15081 [cs.AI] |
| (or arXiv:2610.15081v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.15081 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Silvère Gangloff [view email]
[v1]
Sat, 10 Oct 2026 17:28:23 UTC (51 KB)
来源：arXiv cs.AI · arxiv.org