← Back to all news
arXiv Machine Learning

Bridge Matching Sampler: Scalable Sampling via Generalized Fixed-Point Diffusion Matching

Bridge Matching Sampler (BMS) is a new diffusion-based sampling method that generalizes least-squares matching objectives into fixed-point iterations using Nelson's relation. It learns a stochastic transport map between any prior and target distribution with a single, scalable, and stable objective, and a damped variant adds regularization to reduce mode collapse. Experiments show BMS can sample at unprecedented scales while preserving mode diversity, achieving state‑of‑the‑art results on complex synthetic densities and high‑dimensional molecular benchmarks.

Related stories

arXiv Machine Learning
Aug 31

Improved off-policy training of diffusion samplers

The paper investigates training diffusion models to sample from distributions defined by unnormalized densities or energy functions. It benchmarks various diffusion-structured inference techniques, including simulation-based variational methods and off-policy approaches such as continuous generative flow networks, highlighting their relative strengths and challenging some prior claims. Additionally, the authors introduce a new exploration strategy for off-policy methods that employs local search in the target space with a replay buffer, demonstrating improved sample quality across multiple target distributions.

By Marcin Sendera, Minsu Kim, Sarthak Mittal, Pablo Lemos, Luca Scimeca, Jarrid Rector-Brooks, Alexandre Adam, Yoshua Bengio, Esmeralda S. Whitammer
diffusionbenchmarks
More like this →
arXiv Machine Learning
Sep 11

Prediction--Loss Alignment for Sampler--Robust Flow Matching Training

The paper examines the common practice in diffusion and flow matching of predicting the clean signal, converting it to a velocity, and training with a velocity-space loss. It shows that this conversion can cause unstable optimization due to singular endpoint amplification, but that prediction–loss alignment removes this source of non‑integrability and ensures a finite second moment for all timesteps, even with uniform sampling. Experiments confirm that aligned objectives remain trainable across different samplers, reconciling theoretical concerns with empirical success.

By Jiadong Hong, Lei Liu, Xinyu Bian, Wenjie Wang, Zhaoyang Zhang
diffusionsafety
More like this →
arXiv Machine Learning
Oct 3

Fixed-point neural samplers on discrete spaces

The paper introduces the Discrete Gibbs Iterative Neural Sampler, a fixed‑point neural sampler designed for discrete, unnormalized distributions. It overcomes mode collapse, provides convergence guarantees, and is not limited to a single reference process, building on masked diffusion and extending to transport between distribution pairs. The method scales to high‑dimensional systems, supports amortized sampling across conditions, and accurately estimates alloy phase diagrams.

By Jiajun He, Denis Blessing, Mouyang Cheng, Yuanqi Du, Carles Domingo-Enrich
diffusion
More like this →
Hugging Face Trending Papers
6d ago

Data Free Training of Diffusion Models for Lattice Field Theories

The paper addresses the computational challenge of sampling equilibrium field configurations in lattice field theory by proposing a data‑free training method for diffusion models. Instead of relying on target distribution samples, the authors train the models using their own generated samples and the target score identity, leveraging importance‑weighted and self‑consistent fixed‑point formulations. Experiments on two‑dimensional φ⁴ theory demonstrate that this approach achieves performance comparable to training with HMC reference data across various phases.

diffusionbenchmarks
More like this →