mlsanityData to 7 Oct 2026

Analog Physical Systems Can Exhibit Double Descent

arXiv 2511.17825 ↗First posted 21 Nov 2025cond-mat.dis-nn cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

An important component of the success of large AI models is double descent, in which networks avoid overfitting as they grow relative to the amount of training data, instead improving their performance on unseen data. Here we demonstrate double descent in a decentralized analog network of self-adjusting resistive elements. This system trains itself and performs tasks without a digital processor, offering potential gains in energy efficiency and speed -- but must endure component non-idealities. We find that standard training fails to yield double descent, but a modified protocol that accommodates this inherent imperfection succeeds. Our findings show that analog physical systems, if appropriately trained, can exhibit behaviors underlying the success of digital AI. Further, they suggest that biological systems might similarly benefit from over-parameterization.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 17 of its 42 references that are in the map. First posted on arXiv 21 Nov 2025.

Builds on

17 of 42 references.

References · 17

  1. 2018Reconciling modern machine learning practice and the bias-variance trade-offcited by 919
  2. 2016Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagationcited by 251
  3. 2021Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systemscited by 125
  4. 2018A jamming transition from under- to over-parametrization affects loss landscape and generalizationcited by 65
  5. 2024Training of Physical Neural Networkscited by 47
  6. 2023Backpropagation-free Training of Deep Physical Neural Networkscited by 42
  7. 2023Double Descent Demystified: Identifying, Interpreting & Ablating the Sources of a Deep Learning Puzzlecited by 36
  8. 2023Training an Ising Machine with Equilibrium Propagationcited by 29
  9. 2020Memorizing without overfitting: Bias, variance, and interpolation in over-parameterized modelscited by 26
  10. 2023Machine Learning Without a Processor: Emergent Learning in a Nonlinear Electronic Metamaterialcited by 19
  11. 2019Algorithm for Training Neural Networks on Resistive Device Arrayscited by 18
  12. 2022Frequency propagation: Multi-mechanism learning in nonlinear physical networkscited by 11
5 more
  1. 2024Analog In-Memory Computing Attention Mechanism for Fast and Energy-Efficient Large Language Modelscited by 11
  2. 2024Arbitrary control over multimode wave propagation for machine learningcited by 9
  3. 2023Multiplexed gradient descent: Fast online training of modern datasets on hardware neural networks without backpropagationcited by 8
  4. 2022Desynchronous Learning in a Physics-Driven Learning Networkcited by 5
  5. 2024Training all-mechanical neural networks for task learning through in situ backpropagationcited by 4

Built on by

Not cited yet.