Fengzhe Zhang

I am an incoming PhD student at the Gatsby Computational Neuroscience Unit, working on machine learning, generative modelling, variational inference, diffusion-based samplers, and probabilistic machine learning.

Previously, I was a Research Assistant at the University of Cambridge, supervised by Prof. Jose Miguel Hernandez-Lobato, where I worked on diffusion-based generative samplers for molecular energy functions. I completed an MPhil in Machine Learning and Machine Intelligence at Cambridge with Distinction, and a BSc in Mathematics at Imperial College London with First Class Honours.

My recent work studies how consistency models and importance sampling can make sampling from Boltzmann distributions more efficient without introducing bias. I am especially interested in methods that make generative models useful as scientific computing tools: faster samplers, better uncertainty handling, and algorithms that connect statistical modelling with physical systems.

news

Dec 01, 2024	First-author work on efficient and unbiased Boltzmann sampling appeared at ML for Physical Sciences at NeurIPS 2024.
Oct 01, 2024	Started as a Research Assistant at the University of Cambridge, working on diffusion-based generative samplers for molecular energy functions.
Sep 01, 2024	Completed the MPhil in Machine Learning and Machine Intelligence at the University of Cambridge with Distinction.

selected publications

ML4PS
Efficient and Unbiased Sampling of Boltzmann Distributions via Consistency Models

Fengzhe Zhang, J. He, L. Midgley, and 2 more authors

In Machine Learning for Physical Sciences Workshop at NeurIPS, 2024

Abs Bib PDF

This work studies efficient, unbiased sampling of Boltzmann distributions by combining consistency models with importance sampling, reducing the number of function evaluations while preserving effective sample size on synthetic and equivariant n-body systems.
@inproceedings{zhang2024efficient, title = {Efficient and Unbiased Sampling of Boltzmann Distributions via Consistency Models}, author = {Zhang, Fengzhe and He, J. and Midgley, L. and Antor{\'a}n, J. and Hern{\'a}ndez-Lobato, Jos{\'e} Miguel}, booktitle = {Machine Learning for Physical Sciences Workshop at NeurIPS}, year = {2024}, }