Ziming Liu

This article examines how neural networks manage resources for complex tasks.

MIPS offers a new way to create programs from neural networks.

This study explores how neural network representations evolve during training, inspired by nature.

Kolmogorov-Arnold Networks offer innovative solutions for data analysis and learning.

New methods help find integrable systems using machine learning and human expertise.

A study on transformers' capabilities in modeling simple harmonic oscillators.

WallFacer improves efficiency in training long sequence Transformer models with optimized communication.

KANs bridge AI and science, simplifying complex problems and enhancing research.

Using VR for better tornado warning communication can save lives.