Statistical Physics Meets Machine Learning II

Presentations

• Statistical Mechanics of Function Vectors

  ORAL

  March 18, 2026, 12:00 PM – March 18, 2026, 12:12 PM

  Presenters

  • Ravin Raj

    • Princeton University

  Authors

  • Ravin Raj

    • Princeton University

  • Gautam Reddy

    • Princeton University

  View abstract →

• Adaptively Guided Latent Diffusion for Time-Varying Inverse Problems in Particle Accelerators

  ORAL

  March 18, 2026, 12:12 PM – March 18, 2026, 12:24 PM

  Publication: A. Scheinker and A. Williams. "Latent diffusion can map beam loss to two-dimensional phase-space projections." Physical Review Accelerators and Beams 28.9 (2025): 094602. https://doi.org/10.1103/rqg9-g3dp

  Presenters

  • Alexander Scheinker

    • Los Alamos National Laboratory (LANL)

  Authors

  • Alexander Scheinker

    • Los Alamos National Laboratory (LANL)

  View abstract →

• Universality of LLM mechanisms across scale and diversity

  ORAL

  March 18, 2026, 12:24 PM – March 18, 2026, 12:36 PM

  Presenters

  • Lindsay Maleckar Smith

    • Princeton University

  Authors

  • Lindsay Maleckar Smith

    • Princeton University

  • Gautam Reddy

    • Princeton University

  • David J Schwab

    • The Graduate Center, City University of New York

  View abstract →

• Conditioning generative models without retraining: How to make a horse into a car

  ORAL

  March 18, 2026, 12:36 PM – March 18, 2026, 12:48 PM

  Presenters

  • Huan Souza

    • Boston University

  Authors

  • Huan Souza

    • Boston University

  • Pankaj Mehta

    • Boston University

  View abstract →

• Machine learning topological defect formation

  ORAL

  March 18, 2026, 12:48 PM – March 18, 2026, 1:00 PM

  Publication: ​​​​​​​arXiv:2508.20347

  Presenters

  • Fumika Suzuki

    • University of Tokyo ICEPP
    • Los Alamos National Laboratory (LANL)

  Authors

  • Fumika Suzuki

    • University of Tokyo ICEPP
    • Los Alamos National Laboratory (LANL)

  • Ying Wai Li

    • Los Alamos National Laboratory

  • Wojciech H Zurek

    • Los Alamos Natl Lab

  View abstract →

• Inferring Low-rank Energy-based Models and the Renormalization Group

  ORAL

  March 18, 2026, 1:00 PM – March 18, 2026, 1:12 PM

  Presenters

  • Cheyne Weis

    • University of Chicago
    • U Chicago

  Authors

  • Cheyne Weis

    • University of Chicago
    • U Chicago

  • Kyle Bojanek

    • University of Chicago

  • Stephanie E Palmer

    • University of Chicago

  • Peter B Littlewood

    • University of Chicago

  View abstract →

• Bidirectional Nonlinear Optical Tomography: Unbiased Characterization of Off- and On-Chip Coupling Efficiencies

  ORAL

  March 18, 2026, 1:12 PM – March 18, 2026, 1:24 PM

  Publication: arXiv:2510.13110

  Presenters

  • Bo-Han Wu

    • University of Hawaii at Manoa

  Authors

  • Bo-Han Wu

    • University of Hawaii at Manoa

  • Mahmoud Jalali Mehrabad

    • Massachusetts Institute of Technology
    • MIT

  • Dirk Englund

    • Massachusetts Institute of Technology
    • MIT

  View abstract →

• Is Grokking a Computational Glass Relaxation?

  ORAL

  March 18, 2026, 1:24 PM – March 18, 2026, 1:36 PM

  Publication: [1] Yang, E., Zhang, X., Shang, Y., & Zhang, G. (2025). High-entropy Advantage in Neural Networks' Generalizability. arXiv:2503.13145. In Revision.
  [2] Zhang, X., Shang, Y., Yang, E., & Zhang, G. (2025). Is Grokking a Computational Glass Relaxation? arXiv:2505.11411. Advances in Neural Information Processing Systems, 2025. (Accepted as Spotlight)

  Presenters

  • Entao Yang

    • Air Liquide USA

  Authors

  • Entao Yang

    • Air Liquide USA

  • Xiaotian Zhang

    • City University of Hong Kong

  • Yue Shang

    • University of Pennsylvania

  • Ge Zhang

    • City Univ of Hong Kong

  View abstract →

• Statistical analysis of neural network structures for the Iris dataset

  ORAL

  March 18, 2026, 1:36 PM – March 18, 2026, 1:48 PM

  Presenters

  • Xiguang Yang

    • University of Georgia

  Authors

  • Xiguang Yang

    • University of Georgia

  • Michael Bachmann

    • University of Georgia

  View abstract →

• Exact solution of the frustrated Potts model with next-nearest-neighbor interactions in one dimension via AI bootstrapping

  ORAL

  March 18, 2026, 1:48 PM – March 18, 2026, 2:00 PM

  Publication: Weiguo Yin, Phys. Rev. B 112, 094424 (2025). DOI: https://doi.org/10.1103/y5vc-3t6q

  Presenters

  • Weiguo Yin

    • Brookhaven National Laboratory (BNL)

  Authors

  • Weiguo Yin

    • Brookhaven National Laboratory (BNL)

  View abstract →

• Semantic Chunking and the Entropy of Natural Language

  ORAL

  March 18, 2026, 2:00 PM – March 18, 2026, 2:12 PM

  Presenters

  • Weishun Zhong

    • Institute for Advanced Study

  Authors

  • Weishun Zhong

    • Institute for Advanced Study

  • Tankut Can

    • Emory University

  • Mikhail Katkov

    • Institute for Advanced Study, Princeton and Weizmann Institute of Science

  • Misha Tsodyks

    • Institute for Advanced Study, Princeton and Weizmann Institute of Science

  View abstract →

• Stochastic Resetting of Reinforcement Learning Agents

  ORAL

  March 18, 2026, 2:12 PM – March 18, 2026, 2:24 PM

  Presenters

  • Jello Zhou

    • Stanford University

  Authors

  • Jello Zhou

    • Stanford University

  • David J Schwab

    • The Graduate Center, City University of New York

  • Wave Ngampruetikorn

    • University of Sydney

  View abstract →

• Expressivity of first-order phase transitions

  ORAL

  March 18, 2026, 2:24 PM – March 18, 2026, 2:36 PM

  Presenters

  • Aditya Gandotra

    • University of Chicago

  Authors

  • Aditya Gandotra

    • University of Chicago

  • Mason N Rouches

    • University of Chicago

  • Constantine G Evans

    • Maynooth University

  • Arvind Murugan

    • University of Chicago

  View abstract →

• Learning Order: Can Neural Networks Discover Phase Transitions Without Symmetry Functions?

  ORAL

  March 18, 2026, 2:36 PM – March 18, 2026, 2:48 PM

  Publication: [1] Carina Karner "Variational autoencoders can detect phase signatures from raw trajectory data", in preparation, 2026

  Presenters

  • Carina Karner

    • Vienna University of Technology

  Authors

  • Carina Karner

    • Vienna University of Technology

  View abstract →

• Abstract Withdrawn

  ORAL · Withdrawn

  March 18, 2026, 2:48 PM – March 18, 2026, 3:00 PM

  View abstract →