Effective Test-Time Scaling of Discrete Diffusion through Iterative Refinement

• URL: http://arxiv.org/abs/2511.05562v1
• Date: Tue, 04 Nov 2025 02:33:23 GMT
• Title: Effective Test-Time Scaling of Discrete Diffusion through Iterative Refinement
• Authors: Sanghyun Lee, Sunwoo Kim, Seungryong Kim, Jongho Park, Dongmin Park,
• Abstract summary: We introduce Iterative Reward-Guided Refinement (IterRef), a novel test-time scaling method tailored to discrete diffusion.<n>We formalize this process within a Multiple-Try Metropolis framework, proving convergence to the reward-aligned distribution.<n>IterRef achieves striking gains under low compute budgets, far surpassing prior state-of-the-art baselines.
• Score: 51.54933696252104
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Test-time scaling through reward-guided generation remains largely unexplored for discrete diffusion models despite its potential as a promising alternative. In this work, we introduce Iterative Reward-Guided Refinement (IterRef), a novel test-time scaling method tailored to discrete diffusion that leverages reward- guided noising-denoising transitions to progressively refine misaligned intermediate states. We formalize this process within a Multiple-Try Metropolis (MTM) framework, proving convergence to the reward-aligned distribution. Unlike prior methods that assume the current state is already aligned with the reward distribution and only guide the subsequent transition, our approach explicitly refines each state in situ, progressively steering it toward the optimal intermediate distribution. Across both text and image domains, we evaluate IterRef on diverse discrete diffusion models and observe consistent improvements in reward-guided generation quality. In particular, IterRef achieves striking gains under low compute budgets, far surpassing prior state-of-the-art baselines.

Related papers

• Inference-Time Alignment for Diffusion Models via Doob's Matching [16.416975860645724]
  Inference-time alignment for diffusion models aims to adapt a pre-trained diffusion model toward a target distribution without retraining the base score network.<n>We introduce Doob's matching, a novel framework for guidance estimation grounded in Doob's $h$-transform.<n>We prove non-asymptotic convergence guarantees for the generated distributions in the 2-Wasserstein distance.
  arXiv  Detail & Related papers  (2026-01-10T10:28:06Z)
• TAG:Tangential Amplifying Guidance for Hallucination-Resistant Diffusion Sampling [53.61290359948953]
  Tangential Amplifying Guidance (TAG) operates solely on trajectory signals without modifying the underlying diffusion model.<n>We formalize this guidance process by leveraging a first-order Taylor expansion.<n> TAG is a plug-and-play, architecture-agnostic module that improves diffusion sampling fidelity with minimal computational addition.
  arXiv  Detail & Related papers  (2025-10-06T06:53:29Z)
• Discrete Guidance Matching: Exact Guidance for Discrete Flow Matching [36.348940136801296]
  A novel guidance framework for discrete data is proposed to address this problem.<n>We derive the exact transition rate for the desired distribution given a learned discrete flow matching model.<n>We demonstrate the effectiveness of our proposed guidance on energy-guided simulations and preference alignment on text-to-image generation and multimodal understanding tasks.
  arXiv  Detail & Related papers  (2025-09-26T05:51:31Z)
• Observation Adaptation via Annealed Importance Resampling for Partially Observable Markov Decision Processes [4.830416359005018]
  Partially observable Markov decision processes (POMDPs) are a general mathematical model for sequential decision-making in environments under state uncertainty.<n>Online solvers typically use bootstrap particle filters based on importance resampling for updating the belief distribution.<n>We propose an approach that constructs a sequence of bridge distributions between the state-transition and optimal distributions through iterative Monte Carlo steps.
  arXiv  Detail & Related papers  (2025-03-25T03:05:00Z)
• Reward-Guided Iterative Refinement in Diffusion Models at Test-Time with Applications to Protein and DNA Design [87.58981407469977]
  We propose a novel framework for inference-time reward optimization with diffusion models inspired by evolutionary algorithms.<n>Our approach employs an iterative refinement process consisting of two steps in each iteration: noising and reward-guided denoising.
  arXiv  Detail & Related papers  (2025-02-20T17:48:45Z)
• RDPM: Solve Diffusion Probabilistic Models via Recurrent Token Prediction [17.005198258689035]
  Diffusion Probabilistic Models (DPMs) have emerged as the de facto approach for high-fidelity image synthesis.<n>We introduce a novel generative framework, the Recurrent Diffusion Probabilistic Model (RDPM), which enhances the diffusion process through a recurrent token prediction mechanism.
  arXiv  Detail & Related papers  (2024-12-24T12:28:19Z)
• Broadening Target Distributions for Accelerated Diffusion Models via a Novel Analysis Approach [49.97755400231656]
  We show that a new accelerated DDPM sampler achieves accelerated performance for three broad distribution classes not considered before.<n>Our results show an improved dependency on the data dimension $d$ among accelerated DDPM type samplers.
  arXiv  Detail & Related papers  (2024-02-21T16:11:47Z)
• Distributed Markov Chain Monte Carlo Sampling based on the Alternating Direction Method of Multipliers [143.6249073384419]
  In this paper, we propose a distributed sampling scheme based on the alternating direction method of multipliers. We provide both theoretical guarantees of our algorithm's convergence and experimental evidence of its superiority to the state-of-the-art. In simulation, we deploy our algorithm on linear and logistic regression tasks and illustrate its fast convergence compared to existing gradient-based methods.
  arXiv  Detail & Related papers  (2024-01-29T02:08:40Z)
• Observation-Guided Diffusion Probabilistic Models [41.749374023639156]
  We propose a novel diffusion-based image generation method called the observation-guided diffusion probabilistic model (OGDM) Our approach reestablishes the training objective by integrating the guidance of the observation process with the Markov chain. We demonstrate the effectiveness of our training algorithm using diverse inference techniques on strong diffusion model baselines.
  arXiv  Detail & Related papers  (2023-10-06T06:29:06Z)
• Adaptive Annealed Importance Sampling with Constant Rate Progress [68.8204255655161]
  Annealed Importance Sampling (AIS) synthesizes weighted samples from an intractable distribution. We propose the Constant Rate AIS algorithm and its efficient implementation for $alpha$-divergences.
  arXiv  Detail & Related papers  (2023-06-27T08:15:28Z)
• Conjugated Discrete Distributions for Distributional Reinforcement Learning [0.0]
  We show that one of the most successful methods may not yield an optimal policy if we have a non-deterministic process. We argue that distributional reinforcement learning lends itself to remedy this situation completely.
  arXiv  Detail & Related papers  (2021-12-14T14:14:49Z)
• Learning Invariant Representations and Risks for Semi-supervised Domain Adaptation [109.73983088432364]
  We propose the first method that aims to simultaneously learn invariant representations and risks under the setting of semi-supervised domain adaptation (Semi-DA) We introduce the LIRR algorithm for jointly textbfLearning textbfInvariant textbfRepresentations and textbfRisks.
  arXiv  Detail & Related papers  (2020-10-09T15:42:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.