Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models

• URL: http://arxiv.org/abs/2512.02044v1
• Date: Wed, 26 Nov 2025 09:49:48 GMT
• Title: Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models
• Authors: Kecheng Chen, Ziru Liu, Xijia Tao, Hui Liu, Xinyu Fu, Suiyun Zhang, Dandan Tu, Lingpeng Kong, Rui Liu, Haoliang Li,
• Abstract summary: Coherent Contextual Decoding (CCD) is a novel inference framework built upon two core innovations.<n>CCD employs a trajectory rectification mechanism that leverages historical context to enhance sequence coherence.<n>Instead of rigid allocations based on diffusion steps, we introduce an adaptive sampling strategy that dynamically adjusts the unmasking budget for each step.
• Score: 64.92045568376705
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Diffusion Language Models (DLMs) have recently achieved significant success due to their any-order generation capabilities. However, existing inference methods typically rely on local, immediate-step metrics such as confidence or entropy which inherently lack a more reliable perspective. This limitation frequently leads to inconsistent sampling trajectories and suboptimal generation quality. To address this, we propose Coherent Contextual Decoding (CCD), a novel inference framework built upon two core innovations. First, CCD employs a trajectory rectification mechanism that leverages historical context to enhance sequence coherence, enabling the early rejection of suboptimal paths. We demonstrate that this mechanism is theoretically equivalent to modeling the consistency of historical steps via the conditional mutual information between context and token predictions. Building on this theoretical insight, we further address the inefficiency of conventional uniform decoding budgets. Instead of rigid allocations based on diffusion steps, we introduce an adaptive sampling strategy that dynamically adjusts the unmasking budget for each step according to our consistency metric. Consequently, our method significantly improves the quality of generation trajectories while accelerating the sampling process. Empirically, our method achieves a simultaneous enhancement in both inference speed and performance across diverse benchmarks on Dream and LLaDA, delivering up to 3.48x speedup alongside 3.91% performance improvement.

Related papers

• Causal Autoregressive Diffusion Language Model [70.7353007255797]
  CARD reformulates the diffusion process within a strictly causal attention mask, enabling dense, per-token supervision in a single forward pass.<n>Our results demonstrate that CARD achieves ARM-level data efficiency while unlocking the latency benefits of parallel generation.
  arXiv  Detail & Related papers  (2026-01-29T17:38:29Z)
• Advancing Reliable Test-Time Adaptation of Vision-Language Models under Visual Variations [67.35596444651037]
  Vision-language models (VLMs) exhibit remarkable zero-shot capabilities but struggle with distribution shifts in downstream tasks when labeled data is unavailable.<n>We propose a Reliable Test-time Adaptation (ReTA) method that enhances reliability from two perspectives.
  arXiv  Detail & Related papers  (2025-07-13T05:37:33Z)
• Unified Enhancement of the Generalization and Robustness of Language Models via Bi-Stage Optimization [2.502393972789905]
  We propose a bi-stage optimization framework to uniformly enhance both the generalization and robustness of LMs.<n>We show that our method significantly improves the generalization and robustness of LMs compared to other existing methods.
  arXiv  Detail & Related papers  (2025-03-19T13:50:36Z)
• 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)
• E2ED^2:Direct Mapping from Noise to Data for Enhanced Diffusion Models [15.270657838960114]
  Diffusion models have established themselves as the de facto primary paradigm in visual generative modeling.<n>We present a novel end-to-end learning paradigm that establishes direct optimization from the final generated samples to initial noises.<n>Our method achieves substantial performance gains in terms of Fr'eche't Inception Distance (FID) and CLIP score, even with fewer sampling steps.
  arXiv  Detail & Related papers  (2024-12-30T16:06:31Z)
• Speculative Diffusion Decoding: Accelerating Language Generation through Diffusion [55.0194604505437]
  Speculative decoding has emerged as a widely adopted method to accelerate large language model inference.<n>This paper proposes an adaptation of speculative decoding which uses discrete diffusion models to generate draft sequences.
  arXiv  Detail & Related papers  (2024-08-10T21:24:25Z)
• Efficient Text-driven Motion Generation via Latent Consistency Training [21.348658259929053]
  We propose a motion latent consistency training framework (MLCT) to solve nonlinear reverse diffusion trajectories.<n>By combining these enhancements, we achieve stable and consistency training in non-pixel modality and latent representation spaces.
  arXiv  Detail & Related papers  (2024-05-05T02:11:57Z)
• Learn to Optimize Denoising Scores for 3D Generation: A Unified and Improved Diffusion Prior on NeRF and 3D Gaussian Splatting [60.393072253444934]
  We propose a unified framework aimed at enhancing the diffusion priors for 3D generation tasks. We identify a divergence between the diffusion priors and the training procedures of diffusion models that substantially impairs the quality of 3D generation.
  arXiv  Detail & Related papers  (2023-12-08T03:55:34Z)
• Federated Distributionally Robust Optimization with Non-Convex Objectives: Algorithm and Analysis [21.913563167426872]
  Asynchronous distributed algorithm named Asynchronous Single-looP alternatIve gRadient projEction is proposed.<n>New uncertainty set, i.e., constrained D-norm uncertainty set, is developed to leverage the prior distribution and flexibly control the degree of robustness.<n> empirical studies on real-world datasets demonstrate that the proposed method can not only achieve fast convergence, but also remain robust against data as well as malicious attacks.
  arXiv  Detail & Related papers  (2023-07-25T01:56:57Z)
• Conditional Denoising Diffusion for Sequential Recommendation [62.127862728308045]
  Two prominent generative models, Generative Adversarial Networks (GANs) and Variational AutoEncoders (VAEs) GANs suffer from unstable optimization, while VAEs are prone to posterior collapse and over-smoothed generations. We present a conditional denoising diffusion model, which includes a sequence encoder, a cross-attentive denoising decoder, and a step-wise diffuser.
  arXiv  Detail & Related papers  (2023-04-22T15:32:59Z)
• Distributed Distributionally Robust Optimization with Non-Convex Objectives [24.64654924173679]
  Asynchronous distributed algorithm named Asynchronous Single-looP alternatIve gRadient projEction is proposed. New uncertainty set, i.e., constrained D-norm uncertainty set, is developed to leverage the prior distribution and flexibly control the degree of robustness. empirical studies on real-world datasets demonstrate that the proposed method can not only achieve fast convergence, but also remain robust against data as well as malicious attacks.
  arXiv  Detail & Related papers  (2022-10-14T07:39:13Z)

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.