arXiv AI
Jul 21
Thermodynamic Limits of Physical Intelligence
arXiv:2602. 05463v2 Announce Type: replace-cross Abstract: Modern AI systems achieve remarkable capabilities at the cost of substantial energy consumption.
Thermodynamic cost of inference and learning in physical neural networks
The paper investigates the thermodynamic cost of inference and learning in physical neural networks. It shows that quasi‑static inference requires no work, while finite‑speed inference incurs work bounded by the Wasserstein‑2 distance between thermal states, roughly $k_B T$ per dimension of the widest layer. Learning, however, has an irreducible cost of a few $k_B T$ per parameter, independent of speed, indicating that memory dominates the thermodynamic price.
Related stories
arXiv Machine Learning
Jul 30
On the robustness of noisy solutions in non-convex neural networks
arXiv:2607. 27000v1 Announce Type: cross Abstract: Optimization in non-convex neural network models is strongly influenced by the geometry of the solution space: sparse, isolated, point-like clusters are typically algorithmically inaccessible, whereas wide and flat regions can be found efficiently despite being relatively rare.
arXiv AI
Jul 2
Scaling Up Thermodynamic AI Models
arXiv:2607. 00170v1 Announce Type: cross 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.
Hugging Face Trending Papers
Jun 30
Scaling Up Thermodynamic AI Models
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.
arXiv Machine Learning
Aug 18
DeepOHeat-v2: Self-Improving Operator Learning for Fast and Trustworthy Thermal Optimization in 3D-IC Design
arXiv:2608. 16080v1 Announce Type: new Abstract: Thermal-aware optimization of multi-die 3D integrated circuits evaluates many designs, each a costly heat-equation solve.
arXiv Machine Learning
Aug 4
Conservation laws determine what physical learning remembers
arXiv:2608. 00097v1 Announce Type: cross Abstract: Physical learning rules such as equilibrium propagation (EP), coupled learning (CL), and adjoint coupled learning (AL) train resistive networks through local measurements.
arXiv Machine Learning
Jun 9
GENERIC-FNO: Embedding Energy Conservation and Entropy Production into Fourier Neural Operators
arXiv:2606. 08343v1 Announce Type: new Abstract: We introduce GENERIC-FNO, the first neural operator to embed the full GENERIC (metriplectic) structure of nonequilibrium thermodynamics -- reversible, energy-conserving dynamics and irreversible, entropy-producing dynamics coupled through the degeneracy conditions -- directly in function space.
arXiv AI
Jun 2
Equilibrium Propagation for Non-Conservative Systems
arXiv:2602. 03670v2 Announce Type: replace-cross Abstract: Equilibrium Propagation (EP) is a physics-inspired learning algorithm that uses stationary states of a dynamical system both for inference and learning.
arXiv Machine Learning
Aug 4
Do Neural Networks Really Beat the Curse of Dimensionality? A Bit-Complexity View
arXiv:2608. 01357v1 Announce Type: new Abstract: Traditional approximation theory measures convergence rates in terms of the number of parameters or degrees of freedom.
arXiv Machine Learning
Jun 30
Anti-Collapse Dynamics and the Emergence of Multi-Time-Scale Learning in Recurrent Neural Networks
arXiv:2606. 29519v1 Announce Type: new Abstract: Long-range learning is hard for recurrent networks trained with stochastic gradient descent, because the influence of a past input fades with the lag $\ell$, and if it fades too fast the dependence cannot be learned from finite data.
Hugging Face Trending Papers
Jun 28
Anti-Collapse Dynamics and the Emergence of Multi-Time-Scale Learning in Recurrent Neural Networks
Long-range learning is hard for recurrent networks trained with stochastic gradient descent, because the influence of a past input fades with the lag $\ell$, and if it fades too fast the dependence cannot be learned from finite data. This fade is captured by an envelope $f(\ell)$.
arXiv Machine Learning
Sep 15
Beyond Noise: Understanding and Overcoming Temperature Effects in Analog DNN Inference
The paper investigates how temperature affects analog deep neural network (DNN) inference, focusing on both stochastic and systematic non‑idealities in analog hardware. Experiments show that temperature‑induced performance loss is mainly driven by systematic errors rather than random noise. The study evaluates various mitigation techniques, finding that noise‑aware training and temperature‑aware calibration—especially hardware‑in‑the‑loop training—best preserve inference accuracy across different thermal conditions.