Independent Metropolis-Hastings

Follow this topic Ask about this topic

Latest papers

8 papers2 papers in the last four weeks, against 2 the four weeks before. 0.0% of all new papers.

Spend Where It Matters: Particle Search with MH Refinement for Budget-Constrained Audio-Visual Alignment
Oct 8, 2026Shivanshu Shekhar, Raghavendra Addanki, Tong Zhang +1Independent Metropolis-Hastings Diffusion Model Alignment
University of Illinois Urbana-Champaign · Adobe Research
Metropolis-Hastings Dominates Importance Resampling for Policy Composition
Oct 2, 2026Alexey Kurennoy, Ramil Yarullin, Fergal ReidIndependent Metropolis-Hastings Language Model Decoding
GRPO-QPS: Target-Preserving Reinforcement Learning for Quantum Posterior Sampling
Sep 13, 2026Yufeng Wang, Parivesh Priye, Lu Wei +1Independent Metropolis-Hastings Markov Chain Monte Carlo
Stony Brook University · Georgia Institute of Technology · Westlake University
Exact Global MCMC with Denoising Diffusion
Aug 31, 2026Mitch HillIndependent Metropolis-Hastings Markov Chain Monte Carlo
Optimal scaling of MCMC algorithms: the Hamiltonian approach
Jul 1, 2026P. Dobson, J. M. Sanz-Serna, K. C. ZygalakisIndependent Metropolis-Hastings Markov Chain Monte Carlo
Maxwell Institute for Mathematical Sciences and Mathematics Department Heriot-Watt University, Edinburgh, EH14 4AS, UK · Departamento de Matem´aticas, Universidad Carlos III de Madrid Avenida Universidad 30, 28911 Legan´es, Madrid · Maxwell Institute for Mathematical Sciences and School of Mathematics University of Edinburgh, Peter Guthrie Tait Rd, EH9 3FD, Edinburgh
The Normalized Maximum Likelihood for Regular Non-Smooth Models: Measure-Theoretic Foundations and Geometric Sampling
May 23, 2026Trenton Lau, Gary P. T. ChoiIndependent Metropolis-Hastings Markov Chain Monte Carlo
Department of Mathematics, The Chinese University of Hong Kong, Hong Kong
Proximal-IMH: Proximal Posterior Proposals for Independent Metropolis-Hastings with Approximate Operators
Feb 24, 2026Youguang Chen, George BirosIndependent Metropolis-Hastings Posterior Sampling
Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, Texas, USA
Variance Reduction for Independent Metropolis
Jun 25, 2024Siran Liu, Petros Dellaportas, Michalis K. TitsiasIndependent Metropolis-Hastings Markov Chain Monte Carlo
Department of Statistical Science, University College London, UK. · Department of Statistics, Athens University of Economics and Business, Greece. · Google DeepMind, UK.

Oct 8, 2026J/K move · Enter open · S save

Spend Where It Matters: Particle Search with MH Refinement for Budget-Constrained Audio-Visual Alignment

Shivanshu Shekhar, Raghavendra Addanki, Tong Zhang, Uttaran Bhattacharya

University of Illinois Urbana-Champaign · Adobe Research

Independent Metropolis-Hastings Diffusion Model Alignment Inference-Time Guidance Audio-Visual Alignment

Open paper AskPDF arXiv

Reward-guided inference-time alignment can improve joint audio-visual generation, but optimizing multiple alignment objectives requires many expensive model evaluations, making existing methods costly under limited compute. Particle-based search enables global exploration but incurs estimation error with a finite population, while Metropolis-Hastings (MH) can refine candidates locally but benefits from a strong initialization. We introduce Spend Where It Matters (\modelname), a two-stage, training-free, reward-agnostic framework that combines particle search with MH refinement under a fixed compute budget. In the first stage, we adaptively allocate particles across denoising checkpoints using online variance estimates, minimizing estimation variance while concentrating compute where it is most useful. This search provides a promising candidate for MH refinement, which uses local noise-and-denoise proposals on completed outputs to further improve reward. By starting from a strong candidate, MH can recover additional reward without requiring a larger particle population or costly global exploration. We show that, at matched compute, combining particle search with MH refinement achieves higher rewards than either approach alone. On JavisBench-Mini, at closely matched compute, \modelname achieves mean relative gains across seven metrics of

30.7\%

over BoN and

14.1\%

over EvoSearch with JavisDiT++, and

9.2\%

over BoN and

17.0\%

over EvoSearch with LTX-2. Beyond joint audio-visual generation, we also demonstrate the benefits of adaptive particle allocation on SD1.5 image generation, improving AestheticScore by

2.2\%

and ImageReward by

6.0\%

over uniform allocation.