Diffusion as a Training Curriculum for Timestep-Free Iterative Reasoning

• URL: http://arxiv.org/abs/2609.01449v1
• Date: Tue, 01 Sep 2026 15:55:01 GMT
• Title: Diffusion as a Training Curriculum for Timestep-Free Iterative Reasoning
• Abstract summary: We introduce a hidden state diffusion denoiser that can be run to arbitrary depth.<n>We reach 99.90% exact solve on Sudoku-Extreme and 98.93% on Maze-Unique.<n>We find that diffusion's primary contribution to our anytime solver is not a sampling procedure at inference, but a denoising training curriculum.
• Score: 1.16829782470264
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Diffusion models and recursive reasoners are both iterative, but they carry information across iterations differently. We add a persistent hidden state to a diffusion denoiser and remove its timestep conditioning, leaving a single shared update that can be run to arbitrary depth. The result is an anytime solver: accuracy keeps improving with inference depth far beyond the rollout lengths and backpropagation window used in training, reaching 99.90% exact solve on Sudoku-Extreme. We also obtain 98.93% solve rate on Maze-Unique. Surprisingly, progressive denoising is unnecessary at inference: holding corruption at its maximum by replacing every non-clue variable with fresh Gaussian noise at each step retains near-perfect solving and converges to stable solutions. This simple noise-injection mechanism enables a single trajectory to efficiently explore the solution space and settle on the correct answer without parallel rollouts, candidate selection, or external verifiers required by prior reasoning models. Nonetheless, ordered annealed corruption remains critical during training, which suggests that diffusion's primary contribution to our anytime solver is not a sampling procedure at inference, but a denoising training curriculum.

Related papers

• Thinking with Looped Flows [19.42368401391554]
  Humans and machines often solve harder problems by spending more time on computation.<n>In deep learning, looped models implement this idea during inference by recurrently updating a hidden state.<n>We propose looped flows, an approach that sidesteps this issue by training the recurrence with local denoising objectives.
  arXiv  Detail & Related papers  (2026-09-10T16:52:54Z)
• Let It Go or Learn to Self-Correct: Continuous Diffusion for Constrained Discrete Tasks [16.76819088768813]
  Denoising Diffusion Probabilistic Models generate samples by repeatedly denoising while keeping each update close to the current noisy state.<n>We show that standard diffusion sampling may preserve early mistakes, even when the model's clean predictions are informative.<n>We introduce self-correction training, which exposes the model to its own predictions, improving robustness to errors that arise during inference.
  arXiv  Detail & Related papers  (2026-09-08T16:45:52Z)
• The Equilibrium Is the Initialization: Lazy Identity Collapse in Physics-Structured Deep Equilibrium Reasoning [0.0]
  Deep equilibrium models promise input-adaptive implicit computation.<n>We report a cautionary study of a port-Hamiltonian DEQ with learned reasoning tasks.<n>We show that the standard zeroing ablation is confounded and gives wildly seed-dependent answers.
  arXiv  Detail & Related papers  (2026-07-13T05:45:35Z)
• What Does a Discrete Diffusion Model Learn? [71.03603607324338]
  discrete diffusion model learns: a denoiser, a score ratio, or a bridge plug-in predictor.<n>We start with a rigorous derivation of the continuous-time Markov chain (CTMC) ELBO for any noising process.<n>Every identity is verified numerically, without approximation, on an exactly solvable model.
  arXiv  Detail & Related papers  (2026-07-06T17:56:11Z)
• Guess & Guide: Gradient-Free Zero-Shot Diffusion Guidance [29.644504577943696]
  We introduce a lightweight likelihood surrogate that eliminates the need to calculate gradients through the denoiser network.<n>Experiments confirm that using our method, the inference cost drops dramatically.<n>At the same time, our approach delivers the highest results in multiple tasks.
  arXiv  Detail & Related papers  (2026-03-09T00:21:53Z)
• Test-Time Scaling with Diffusion Language Models via Reward-Guided Stitching [66.39914384073145]
  We propose a self-consistency framework that turns cheap diffusion-sampled reasoning into a reusable pool of step-level candidates.<n>We find that step-level recombination is most beneficial on harder problems.<n>Our training-free framework improves average accuracy by up to 2 across six math and coding tasks.
  arXiv  Detail & Related papers  (2026-02-26T11:08:39Z)
• Rejection Mixing: Fast Semantic Propagation of Mask Tokens for Efficient DLLM Inference [58.189320101488725]
  DLLMs promise fast non-autoregressive inference but suffer a severe quality-speed trade-off in parallel decoding.<n>We address this by integrating continuous representations into the discrete decoding process, as they preserve rich inter-position dependency.<n>We propose ReMix, a framework that introduces a novel Continuous Mixing State as an intermediate between the initial masked state and the final decoded token state.
  arXiv  Detail & Related papers  (2026-02-26T11:08:11Z)
• Self-diffusion for Solving Inverse Problems [3.8870795921263728]
  We propose self-diffusion, a novel framework for solving inverse problems without relying on pretrained generative models.<n>Self-diffusion exploits the spectral bias of neural networks and modulates it through a scheduled noise process.<n>We demonstrate the effectiveness of our approach on a variety of linear inverse problems, showing that self-diffusion achieves competitive or superior performance compared to other methods.
  arXiv  Detail & Related papers  (2025-10-24T12:57:22Z)
• Solving Inverse Problems using Diffusion with Iterative Colored Renoising [13.203844370996967]
  We show that the approximations produced by existing methods are relatively poor, especially early in the reverse process.<n>We propose a new approach that iteratively reestimates and "renoises" the estimate several times per diffusion step.<n>This iterative approach, which we call Fast Iterative REnoising (FIRE), injects colored noise that is shaped to ensure that the pre-trained diffusion model always sees white noise.
  arXiv  Detail & Related papers  (2025-01-29T08:20:05Z)
• Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment [56.609042046176555]
  suboptimal noise-data mapping leads to slow training of diffusion models. Drawing inspiration from the immiscibility phenomenon in physics, we propose Immiscible Diffusion. Our approach is remarkably simple, requiring only one line of code to restrict the diffuse-able area for each image.
  arXiv  Detail & Related papers  (2024-06-18T06:20:42Z)
• Towards More Accurate Diffusion Model Acceleration with A Timestep Tuner [112.99126045081046]
  A diffusion model, which is formulated to produce an image using thousands of denoising steps, usually suffers from a slow inference speed.<n>We propose a textbftimestep tuner that helps find a more accurate integral direction for a particular interval at the minimum cost.<n>Experiments show that our plug-in design can be trained efficiently and boost the inference performance of various state-of-the-art acceleration methods.
  arXiv  Detail & Related papers  (2023-10-14T02:19:07Z)
• DiffTAD: Temporal Action Detection with Proposal Denoising Diffusion [137.8749239614528]
  We propose a new formulation of temporal action detection (TAD) with denoising diffusion, DiffTAD. Taking as input random temporal proposals, it can yield action proposals accurately given an untrimmed long video.
  arXiv  Detail & Related papers  (2023-03-27T00:40:52Z)
• Consistent Diffusion Models: Mitigating Sampling Drift by Learning to be Consistent [97.64313409741614]
  We propose to enforce a emphconsistency property which states that predictions of the model on its own generated data are consistent across time. We show that our novel training objective yields state-of-the-art results for conditional and unconditional generation in CIFAR-10 and baseline improvements in AFHQ and FFHQ.
  arXiv  Detail & Related papers  (2023-02-17T18:45:04Z)
• Distribution Mismatch Correction for Improved Robustness in Deep Neural Networks [86.42889611784855]
  normalization methods increase the vulnerability with respect to noise and input corruptions. We propose an unsupervised non-parametric distribution correction method that adapts the activation distribution of each layer. In our experiments, we empirically show that the proposed method effectively reduces the impact of intense image corruptions.
  arXiv  Detail & Related papers  (2021-10-05T11:36:25Z)

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.