CRAM 能量模型训练推理速度比 Looped 模型快 1000 倍
Classical Energies Learn to Reason 1,000x Faster than Looped Models
这篇论文用 80 年代的 Hopfield 能量函数改出一个 CRAM，42ms 训练就能解出 99.996% 的 Sudoku-Extreme，比 looped 模型快 1000 倍、推理省 282 倍 FLOPs。做推理模型研究的人应该看看，还能顺手对 Sudoku 基准祛魅。
一篇 arXiv 论文提出 CRAM，一个基于 Hopfield & Tank 上世纪 80 年代能量函数的学习版本，用于训练推理能力。CRAM 仅在一块 GPU 上训练 42 ms 和 132 ms，就在 Sudoku-Extreme 测试集上达到 99.996% 解题率、Maze-Unique 上达到 100%，比 looped 模型至少快 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.
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include CRAM, Hopfield, 推理模型, 能量模型.
Source attribution: curated from arXiv cs.LG (@Benjamin Hoover, Hugh Blayney, James C. R. Whittington).