Energy-guided Recursive Model
2026-07-11 • Machine Learning
Machine Learning
AI summaryⓘ
The authors developed a new approach called the Energy-guided Recursive Model (ERM) to improve how recursive reasoning models pick the best solutions during problem solving. Instead of using less clear methods like voting, ERM uses a mathematical concept called Hopfield energies to score and select the best options. This helps the model work better on tasks like Sudoku, puzzles, and mazes compared to previous methods. Their work shows that adding explicit energy-based criteria can make these reasoning models more reliable and efficient.
Recursive reasoningHopfield networksEnergy functionsParallel temperingInference mechanismsSudoku solvingNeural networksSampling efficiencyProbabilistic modelsOptimization
Authors
Yifei Zhao, Ying Tang
Abstract
Recursive reasoning models address structured problems by repeatedly updating latent states of small neural networks. However, their test-time scaling lacks a principled inference mechanism: increasing depth or stochastic breadth generates more trajectories without a clear criterion for selection, and existing methods predominantly rely on additional q-heads or heuristic voting. Here, we develop the Energy-guided Recursive Model (ERM), which introduces an intrinsic selection principle based on explicit Hopfield energies. ERM leverages Hopfield-type memories of valid local or global structures to define the selector over candidate trajectories. The resulting energy seamlessly integrates with energy-based techniques such as parallel tempering to enhance sampling efficiency and ranking. With $D=64$ recurrent steps and $K=128$ candidates, ERM reaches optimal solutions on Sudoku ($98.97\%$), Pencil Puzzle Bench (PPBench, $88.04\%$) and Maze ($99.30\%$), improving upon recent Probabilistic Tiny Recursive Model and Equilibrium Reasoners. These results suggest that incorporating explicit energy functions into recursive reasoning offers a principled path toward more effective inference.