What Does a Discrete Diffusion Model Learn?

• URL: http://arxiv.org/abs/2607.05381v1
• Date: Mon, 06 Jul 2026 17:56:11 GMT
• Title: What Does a Discrete Diffusion Model Learn?
• Abstract summary: 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.
• Score: 71.03603607324338
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: What does a discrete diffusion model learn: a denoiser, a score ratio, or a bridge plug-in predictor? At the level of jump rates, these are one object in different coordinates, and reading a neural network in the wrong coordinate changes the process being trained and sampled. Starting with a rigorous derivation of the continuous-time Markov chain (CTMC) ELBO for any noising process, boundary terms included, we prove the \emph{Oracle Distance} theorem: the negative ELBO is exactly equal to the data entropy plus the path KL from the oracle reverse process to the learned one, not merely a bound. Its unique optimizer is therefore the conditional expectation of the true reverse jump rate given the current noisy state, and its irreducible cost is the rate at which the forward process $Z_t$ destroys information about the clean data $Z_0$, $-\tfrac{d}{dt}I(Z_0; Z_t)$, so every noising process shares the same best achievable negative ELBO: the data entropy. For sequences with token-factorizing noise, the oracle projection yields three exact coordinates for the optimizer: denoiser, cavity (bridge plug-in), and score, with closed-form conversions among them. This framework identifies which law each loss in the literature actually optimizes, recovering MDM, UDM, SEDD, and GIDD as special cases; explains why denoiser and cavity coincide for masked diffusion but not for uniform diffusion; proves that a denoiser parameterization makes the uniform ELBO diverge at initialization while the bridge plug-in stays finite; and calibrates ELBO implementations exactly at initialization. Every identity is verified numerically, without approximation, on an exactly solvable model.

Related papers

• Provably adaptive sampling with uniform and remasking discrete diffusion models [9.180350432640912]
  We study a first-order sampler based on the leave-one-out denoiser for uniform and remasking processes.<n>Our main result establishes an adaptive sampling guarantee.<n> Numerical experiments on structured synthetic distributions illustrate the predicted dimension-adaptive behavior.
  arXiv  Detail & Related papers  (2026-08-24T17:54:51Z)
• Simplex Relaxation for Discrete Diffusion [58.33696160279345]
  We introduce Simplax, an exact Dirichlet--categorical augmentation that couples each corrupted categorical state with an auxiliary simplex-valued variable.<n>Simplax improves the generative perplexity--entropy tradeoff on unconditional OpenWebText generation.
  arXiv  Detail & Related papers  (2026-08-11T07:58:17Z)
• Uniform Diffusion Models Revisited: Leave-One-Out Denoiser and Absorbing State Reformulation [72.05592785529404]
  We show that the standard plug-in bridge parameterization for UDM is not optimized by the denoising posterior.<n>We also introduce an absorbing-state reformulation of uniform diffusion that preserves the UDM joint law while decomposing it into masked-diffusion-like sampling operations.
  arXiv  Detail & Related papers  (2026-05-21T17:27:19Z)
• Sampling from Flow Language Models via Marginal-Conditioned Bridges [1.43163477873944]
  Flow Language Models (FLMs) adapt continuous flow matching for one-hot encoded token sequences.<n>We argue that the natural sampler for an FLM is instead posterior-predictive.<n>We show that the induced one-step bridge kernel preserves all token-wise posterior-predictive marginals.
  arXiv  Detail & Related papers  (2026-05-13T15:38:48Z)
• Divergence is Uncertainty: A Closed-Form Posterior Covariance for Flow Matching [8.603039700922809]
  Flow matching has become a leading framework for generative modeling, but quantifying the uncertainty of its samples remains an open problem.<n>We show that none of these trade-offs is necessary.<n>Experiments on MNIST confirm that the resulting per-pixel uncertainty maps are semantically meaningful.
  arXiv  Detail & Related papers  (2026-05-01T04:25:00Z)
• Efficient Sampling with Discrete Diffusion Models: Sharp and Adaptive Guarantees [9.180350432640912]
  We study the sampling efficiency of score-based discrete diffusion models under a continuous-time Markov chain (CTMC) formulation.<n>For uniform discrete diffusion, we show that the $$-leaping algorithm achieves an complexity of order $tilde O(d/varepsilon)$.<n>For masking discrete diffusion, we introduce a modified $$-leaping sampler whose convergence rate is governed by an intrinsic information-theoretic quantity.
  arXiv  Detail & Related papers  (2026-02-16T18:48:17Z)
• Continuously Augmented Discrete Diffusion model for Categorical Generative Modeling [87.34677262370924]
  Standard discrete diffusion models treat all unobserved states identically by mapping them to an absorbing [MASK] token.<n>This creates an 'information void' where semantic information that could be inferred from unmasked tokens is lost between denoising steps.<n>We introduce Continuously Augmented Discrete Diffusion, a framework that augments the discrete state space with a paired diffusion in a continuous latent space.
  arXiv  Detail & Related papers  (2025-10-01T18:00:56Z)
• Non-asymptotic bounds for forward processes in denoising diffusions: Ornstein-Uhlenbeck is hard to beat [49.1574468325115]
  This paper presents explicit non-asymptotic bounds on the forward diffusion error in total variation (TV)<n>We parametrise multi-modal data distributions in terms of the distance $R$ to their furthest modes and consider forward diffusions with additive and multiplicative noise.
  arXiv  Detail & Related papers  (2024-08-25T10:28:31Z)
• Towards Faster Non-Asymptotic Convergence for Diffusion-Based Generative Models [49.81937966106691]
  We develop a suite of non-asymptotic theory towards understanding the data generation process of diffusion models. In contrast to prior works, our theory is developed based on an elementary yet versatile non-asymptotic approach.
  arXiv  Detail & Related papers  (2023-06-15T16:30:08Z)

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.