CRAM 能量模型训练推理速度比 Looped 模型快 1000 倍

技术机制上，CRAM 通过在能量面上做下降来求解，无需对求解器做反向传播就能快速训练，自带早停机制，计算量随问题难度自适应，且权重可解释，能学出谜题的约束规则。在 Sudoku-Extreme 上，CRAM 以至少少 282 倍的推理 FLOPs 匹配了 looped 基线模型的最高准确率。

作者据此提出两点结论：能量函数是从数据训练推理的一个很有前景的方向；同时 Sudoku 和 Maze 这类可被平凡求解的基准，不适合用来检验 looped 推理技术的可扩展性。

Classical Energies Learn to Reason 1,000x Faster than Looped Models

Looped models have emerged as a prominent paradigm for reasoning, with successes on simple reasoning puzzles such as Sudoku-Extreme and Maze-Unique. However, these models can be unstable, slow to train, and use many parameters and FLOPs to solve simple puzzles. Descent on an energy not only mitigates these issues but also offers advantages: built-in early stopping, compute that adapts to problem difficulty, fast training without differentiating through a solver, and interpretable weights that learn the constraints of the puzzle. Here we demonstrate this with a learned version of Hopfield & Tank's energy from the 80s, which we term CRAM, solves 99.996% of Sudoku-Extreme and 100% of Maze-Unique test puzzles after 42 ms and 132 ms of training on one GPU, at least 1,000x faster than looped models on these tasks. CRAM also matches baseline looped models at their highest accuracy on Sudoku-Extreme using at least 282x fewer inference FLOPs. Our work suggests two things. First, that energies are an extremely promising direction for training reasoning from data. Second, a caution against using trivially solvable benchmarks like Sudoku and Maze to test the scalability of looped reasoning techniques.

本条情报归属 AITOP「论文」核心栏目，重点聚焦北美前沿实验室与硅谷科技巨头。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：CRAM、Hopfield、推理模型、能量模型，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。

信源核验说明：本文原始线索采编自「arXiv cs.LG」（@Benjamin Hoover, Hugh Blayney, James C. R. Whittington），经 AITOP 多维评分管线结构化提炼。