Scaling Up Thermodynamic AI Models
2026-06-30 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors developed a new training method for AI models that run on special low-power hardware inspired by physics, called Ising machines. They showed that these models can recognize images well, achieving strong accuracy on standard image datasets. They also created and tested a theory that links how much time the machine spends thinking with how accurate it gets. Finally, they explored how to balance speed and accuracy during inference and discussed what this means for building future hardware. Overall, their work helps make these physics-based AI devices more practical for real-world use.
Ising modelthermodynamic computingGibbs samplingbackpropagationconvolutional neural networksimage classificationCIFAR-10inference costautocorrelation timesedge computing
Authors
Andrew G. Moore
Abstract
Thermodynamic computing devices based on the Ising model show great promise for low-power AI inference and edge computing, but scalable methods for training large models for such hardware remain limited. Prior theory shows that the time-averaged behavior of high-temperature Gibbs-sampled Ising systems can implement feed-forward neural inference. We turn this theoretical correspondence into a scalable and purely backpropagation-based algorithm for training deep convolutional networks for thermodynamic inference on Ising machine hardware. Our image classification models achieve accuracies of 94.9% on CIFAR-10 and 76.0% on CIFAR-100 under binary Gibbs sampling. We then develop and experimentally validate a mathematical theory relating inference cost to accuracy and controlling autocorrelation times. Subsequently, we calculate asymptotic results showing that inference cost is bounded by a well-controlled tradeoff with performance and exhibit algorithms for computing optimal inference schedules. Finally, we discuss implications for hardware development and the future of high-temperature thermodynamic AI models.