arXiv Machine Learning

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.

arXiv Machine Learning

Aug 31

Amortizing intractable inference in diffusion models for vision, language, and control

The paper introduces a data‑free learning objective called relative trajectory balance for training diffusion models to sample from a posterior defined by a diffusion prior and an arbitrary black‑box constraint or likelihood. It proves asymptotic correctness of this objective and demonstrates its use across vision, language, and multimodal tasks, including classifier guidance, language infilling, and text‑to‑image generation. Additionally, the method is applied to continuous control with a score‑based behavior prior, achieving state‑of‑the‑art results in offline reinforcement learning.

By Siddarth Venkatraman, Moksh Jain, Luca Scimeca, Minsu Kim, Marcin Sendera, Mohsin Hasan, Luke Rowe, Sarthak Mittal, Pablo Lemos, Emmanuel Bengio, Alexandre Adam, Jarrid Rector-Brooks, Yoshua Bengio, Glen Berseth, Esmeralda S. Whitammer

llms diffusion reinforcement-learning multimodal benchmarks

arXiv AI

Jul 17

Integration Matters: Rollout-Based Training for Constrained Diffusion Models

arXiv:2607. 14398v1 Announce Type: cross Abstract: Constrained generative models aim to produce samples that satisfy complex feasibility constraints while remaining faithful to the data distribution.

By Xiaoxuan Liang, Saeid Naderiparizi, Berend Zwartsenberg, Frank Wood

diffusion fine-tuning

arXiv Machine Learning

Jun 16

Temporal Difference Learning for Diffusion Models

arXiv:2606. 15048v1 Announce Type: new Abstract: Diffusion models are typically trained with objectives that focus on local denoising targets at individual time steps (or adjacent pairs), which do not enforce consistency between predictions along the denoising trajectory.

By Qizhen Ying, Yangchen Pan, Victor Adrian Prisacariu, Junfeng Wen

diffusion reinforcement-learning benchmarks

arXiv Machine Learning

Jun 2

Do Heavy Tails Help Diffusion? On the Subtle Trade-off Between Initialization and Training

arXiv:2605. 13175v2 Announce Type: replace Abstract: Recent works have proposed incorporating heavy-tailed (HT) noise into diffusion- and flow-based generative models, with the goals of better recovering the tails of target distributions and improving generative diversity.

By Hamza Cherkaoui, H\'el\`ene Halconruy, Antonio Ocello

diffusion

arXiv Machine Learning

Jul 28

Learning Sampling Parameters for Diffusion Models

arXiv:2607. 23488v1 Announce Type: new Abstract: Text-to-image diffusion models expose many inference-time sampling parameters, including prompts, negative prompts, classifier-free guidance scales, and noise schedules.

By Arisrei Lim, Yossi Gandelsman

llms diffusion reinforcement-learning multimodal

arXiv AI

Jun 2

Efficient Weighted Sampling via Score-based Generative Models

arXiv:2502. 04646v2 Announce Type: replace-cross Abstract: Weighted sampling -- sampling from a probability density function (PDF) proportional to the product of a base PDF and a weight function -- is a fundamental technique with wide-ranging applications in variance reduction, biased sampling, data augmentation, and more.

By Heasung Kim, Taekyun Lee, Hyeji Kim, Gustavo de Veciana

diffusion benchmarks

arXiv Machine Learning

Sep 1

Adaptive teachers for amortized samplers

arXiv:2410.01432v3 Announce Type: replace Abstract: Amortized inference is the task of training a parametric model, such as a neural network, to approximate a distribution with a given unnormalized d...

By Minsu Kim, Sanghyeok Choi, Taeyoung Yun, Emmanuel Bengio, Leo Feng, Jarrid Rector-Brooks, Sungsoo Ahn, Jinkyoo Park, Esmeralda S. Whitammer, Yoshua Bengio

diffusion reinforcement-learning

arXiv AI

Jun 16

Proximal Policy Optimization for Amortized Discrete Sampling

arXiv:2606. 15793v1 Announce Type: cross Abstract: This paper explores policy gradient algorithms for training stochastic policies to sample from structured discrete probability distributions under the Generative Flow Network (GFlowNet) framework.

By Anna Zykova-Myzina, Timofei Gritsaev, Daniil Tiapkin, Nikita Morozov

reinforcement-learning benchmarks

arXiv Machine Learning

Jul 9

Diffusion Models in Simulation-Based Inference: A Tutorial Review

arXiv:2512. 20685v3 Announce Type: replace-cross Abstract: Diffusion models have recently emerged as powerful learners for simulation-based inference (SBI), enabling fast and accurate estimation of latent parameters from simulated and real data.

By Jonas Arruda, Niels Bracher, Ullrich K\"othe, Jan Hasenauer, Stefan T. Radev

diffusion

arXiv Machine Learning

Oct 12

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.

By Denis Blessing, Lorenz Richter, Julius Berner, Egor Malitskiy, Gerhard Neumann

diffusion benchmarks

Hugging Face Trending Papers

Jul 26

Learning Sampling Parameters for Diffusion Models

Text-to-image diffusion models expose many inference-time sampling parameters, including prompts, negative prompts, classifier-free guidance scales, and noise schedules. These parameters are typically manually chosen once and then held fixed across prompts and denoising timesteps, even though different prompts and stages of generation can benefit from different parameter values.

llms diffusion reinforcement-learning multimodal

arXiv Machine Learning

Oct 12

Amortised Inference through One-Step Implicit Sampling

The paper introduces OSIS, a one‑step implicit sampler that uses short‑time MCMC dynamics to define a local target for an implicitly modelled distribution. In the asymptotic limit, the model converges to a fixed point of the MCMC kernel, enabling unbiased sampling from the Bayesian posterior. Experiments show OSIS outperforms multi‑step amortised samplers and non‑amortised MCMC baselines while being far more efficient at sampling time.

By Vincent Pauline, Kirill Tamogashev, Arran Carter, Sanghyeok Choi, Stefan Bauer, Esmeralda S. Whitammer

diffusion

Related stories