activity
20242026
collaborators

6 papers

cs.LG2026

Sampling from Flow Language Models via Marginal-Conditioned Bridges

Iskander Azangulov, Leo Zhang

Flow Language Models (FLMs) are a recently introduced class of language models which adapt continuous flow matching for one-hot encoded token sequences. Their denoisers have a spec…

stat.ML2026

Accelerated Parallel Tempering via Neural Transports

Leo Zhang, Peter Potaptchik, Jiajun He +5

Markov Chain Monte Carlo (MCMC) algorithms are essential tools in computational statistics for sampling from unnormalised probability distributions, but can be fragile when targeti…

cs.LG2026

CREPE: Controlling Diffusion with Replica Exchange

Jiajun He, Paul Jeha, Peter Potaptchik +5

Inference-time control of diffusion models aims to steer model outputs to satisfy new constraints without retraining. Previous approaches have mostly relied on heuristic guidance o…

cs.AR2026

Memory-Guided Unified Hardware Accelerator for Mixed-Precision Scientific Computing

Chuanzhen Wang, Leo Zhang, Eric Liu

Recent hardware acceleration advances have enabled powerful specialized accelerators for finite element computations, spiking neural network inference, and sparse tensor operations…

stat.ML2025

The Cosine Schedule is Fisher-Rao-Optimal for Masked Discrete Diffusion Models

Leo Zhang, Saifuddin Syed

In this work, we study the problem of choosing the discretisation schedule for sampling from masked discrete diffusion models in terms of the information geometry of the induced pr…

cs.LG2024

Metric Flow Matching for Smooth Interpolations on the Data Manifold

Kacper Kapuśniak, Peter Potaptchik, Teodora Reu +5

Matching objectives underpin the success of modern generative models and rely on constructing conditional paths that transform a source distribution into a target distribution. Des…