website

Understand what physics/algorithms do transformers learn internally when trained on planetary motion

We study toy models of skill learning.

Getting crystal-like representations with harmonic loss

Kolmogorov Arnold Networks

discovering interpretable conservation laws from differential equations

Brain-Inspired Modular Training (BIMT), a method for making neural networks more modular and interpretable.

Omnigrok: Grokking Beyond Algorithmic Data

AI Poincare 2.0

Machine Learning hidden symmetries

analyze optimizer properties

Code for "Machine-Learning Non-Conservative Dynamics for New-Physics Detection" (arXiv: 2106.00026)

Counting the number of conservation laws from trajectory data

Schrodinger Principal Component Analysis

2019 DATAOPEN held by Correlation One, sponsored by Citadel

Quantum-inspired Hamiltonian Monte Carlo code

expectation maximization

Naive Bayes method

LIU ZIMING's resume

anti mapping of Hydro