Beyond Static Cutoffs: One-Shot Dynamic Thresholding for Diffusion Language Models

• URL: http://arxiv.org/abs/2511.02077v1
• Date: Mon, 03 Nov 2025 21:30:03 GMT
• Title: Beyond Static Cutoffs: One-Shot Dynamic Thresholding for Diffusion Language Models
• Abstract summary: One-Shot Dynamic Thresholding (OSDT) calibrates thresholds on a single sequence and applies them to subsequent inputs with negligible overhead.<n>Our results suggest broader opportunities to leverage reusable task-level confidence signatures for more general-purpose algorithmic and systems innovations in diffusion decoding.
• Score: 3.4202631641965238
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Masked diffusion language models (MDLMs) are becoming competitive with their autoregressive counterparts but typically decode with fixed steps and sequential unmasking. To accelerate decoding, recent work such as Fast-dLLM enables parallel decoding via a static global confidence threshold, yet we observe strong block- and step-wise confidence fluctuations and, within a dataset, near-identical confidence trajectories across inputs as measured by cosine similarity. Motivated by these observations, we introduce One-Shot Dynamic Thresholding (OSDT), which calibrates thresholds on a single sequence and applies them to subsequent inputs with negligible overhead. On GPQA, GSM8K, and HumanEval, OSDT attains superior accuracy-throughput trade-offs (+24% tokens/s on GSM8K at the best accuracy, +45% on GPQA with comparable accuracy, and +50% on HumanEval with a modest accuracy gap). Beyond these results, our findings suggest broader opportunities to leverage reusable task-level confidence signatures for more general-purpose algorithmic and systems innovations in diffusion decoding.

Related papers

• On the Quantization Robustness of Diffusion Language Models in Coding Benchmarks [0.09665867043802821]
  Auto-regressive Large Language Models (LLMs) achieve strong performance on coding tasks, but incur high memory and inference costs.<n>We investigate the application and robustness of PTQ techniques, specifically GPTQ and a modified Hessian-Aware Quantization (HAWQ) algorithm.
  arXiv  Detail & Related papers  (2026-04-22T00:53:43Z)
• Confidence-Based Decoding is Provably Efficient for Diffusion Language Models [8.527157686215117]
  Diffusion language models (DLMs) have emerged as a promising alternative to autoregressive (AR) models for language modeling.<n>We develop the first theoretical analysis framework for confidence-based decoding in DLMs.
  arXiv  Detail & Related papers  (2026-03-23T17:43:21Z)
• Self-Rewarding Sequential Monte Carlo for Masked Diffusion Language Models [58.946955321428845]
  This work presents self-rewarding sequential Monte Carlo (SMC)<n>Our algorithm stems from the observation that most existing MDLMs rely on a confidence-based sampling strategy.<n>We introduce the trajectory-level confidence as a self-rewarding signal for assigning particle importance weights.
  arXiv  Detail & Related papers  (2026-02-02T09:21:45Z)
• STDD:Spatio-Temporal Dynamics-Driven Token Refinement in Diffusion Language Models [12.172699141988728]
  diffusion language models (DLMs) generate text by iteratively denoising all token positions in parallel.<n>We propose a novel remasking approach that dynamically detects Temporal Variance and Spatial Deviance of each token.<n>Our approach significantly improves the operational efficiency of DLMs across mainstream datasets, achieving speedups of up to 8.9 times.
  arXiv  Detail & Related papers  (2025-12-07T12:53:48Z)
• Enhancing Trustworthiness with Mixed Precision: Benchmarks, Opportunities, and Challenges [12.438306093697]
  Large language models (LLMs) have shown promising performance across various tasks.<n>LLMs' autoregressive decoding process poses significant challenges for efficient deployment on existing AI hardware.
  arXiv  Detail & Related papers  (2025-11-27T14:17:43Z)
• From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models [19.97248408121574]
  Diffusion Language Models (DLMs) offer comparable accuracy with faster inference speed via parallel decoding.<n>High-confidence tokens carry negligible information and strictly relying on them limits the effective progress made in each decoding round.<n>We propose Explore-Then-Exploit (ETE), a training-free decoding strategy that maximizes information throughput and decoding efficiency.
  arXiv  Detail & Related papers  (2025-11-26T06:38:37Z)
• Continuous Autoregressive Language Models [56.49239051750678]
  We introduce Continuous Autoregressive Language Models (CALM)<n>CALM uses a high-fidelity autoencoder to compress a chunk of K tokens into a single continuous vector.<n>We develop a comprehensive likelihood-free framework that enables robust training, evaluation, and controllable sampling.
  arXiv  Detail & Related papers  (2025-10-31T17:58:11Z)
• Towards Better Code Generation: Adaptive Decoding with Uncertainty Guidance [42.737012213197865]
  AdaDec is an adaptive decoding framework that employs a lookahead-based, uncertainty-aware pause-and-rerank mechanism.<n>AdaDec achieves up to 20.9% absolute gains in Pass@1 accuracy compared with greedy decoding.<n>By applying reranking only when necessary, AdaDec reduces computational overhead and latency, enhancing efficiency alongside reliability.
  arXiv  Detail & Related papers  (2025-06-10T16:49:46Z)
• Payload-Aware Intrusion Detection with CMAE and Large Language Models [0.0]
  Intrusion Detection Systems (IDS) are crucial for identifying malicious traffic, yet traditional signature-based methods struggle with zero-day attacks and high false positive rates.<n>This study proposes Xavier-CMAE, an enhanced Convolutional Multi-Head Attention Ensemble (CMAE) model that improves detection accuracy while reducing computational overhead.
  arXiv  Detail & Related papers  (2025-03-23T02:56:32Z)
• Breaking Through the Spike: Spike Window Decoding for Accelerated and Precise Automatic Speech Recognition [26.665132884613477]
  The Spike Window Decoding algorithm greatly improves the inference speed by making the number of frames decoded in WFST linearly related to the number of spiking frames in the CTC output.<n>Our method achieves SOTA recognition accuracy with significantly accelerates decoding speed, proven across both AISHELL-1 and large-scale In-House datasets.
  arXiv  Detail & Related papers  (2025-01-01T12:20:07Z)
• Model Inversion Attacks Through Target-Specific Conditional Diffusion Models [54.69008212790426]
  Model inversion attacks (MIAs) aim to reconstruct private images from a target classifier's training set, thereby raising privacy concerns in AI applications. Previous GAN-based MIAs tend to suffer from inferior generative fidelity due to GAN's inherent flaws and biased optimization within latent space. We propose Diffusion-based Model Inversion (Diff-MI) attacks to alleviate these issues.
  arXiv  Detail & Related papers  (2024-07-16T06:38:49Z)
• Masked Thought: Simply Masking Partial Reasoning Steps Can Improve Mathematical Reasoning Learning of Language Models [102.72940700598055]
  In reasoning tasks, even a minor error can cascade into inaccurate results. We develop a method that avoids introducing external resources, relying instead on perturbations to the input. Our training approach randomly masks certain tokens within the chain of thought, a technique we found to be particularly effective for reasoning tasks.
  arXiv  Detail & Related papers  (2024-03-04T16:21:54Z)
• BLSTM-Based Confidence Estimation for End-to-End Speech Recognition [41.423717224691046]
  Confidence estimation is an important function for developing automatic speech recognition (ASR) applications. Recent E2E ASR systems show high performance (e.g., around 5% token error rates) for various ASR tasks. We employ a bidirectional long short-term memory (BLSTM)-based model as a strong binary-class (correct/incorrect) sequence labeler.
  arXiv  Detail & Related papers  (2023-12-22T11:12:45Z)
• Perception and Semantic Aware Regularization for Sequential Confidence Calibration [12.265757315192497]
  We propose a Perception and Semantic aware Sequence Regularization framework. We introduce a semantic context-free recognition and a language model to acquire similar sequences with high perceptive similarities and semantic correlation. Experiments on canonical sequence recognition tasks, including scene text and speech recognition, demonstrate that our method sets novel state-of-the-art results.
  arXiv  Detail & Related papers  (2023-05-31T02:16:29Z)
• Pretraining Without Attention [114.99187017618408]
  This work explores pretraining without attention by using recent advances in sequence routing based on state-space models (SSMs) BiGS is able to match BERT pretraining accuracy on GLUE and can be extended to long-form pretraining of 4096 tokens without approximation.
  arXiv  Detail & Related papers  (2022-12-20T18:50:08Z)
• Fast Entropy-Based Methods of Word-Level Confidence Estimation for End-To-End Automatic Speech Recognition [86.21889574126878]
  We show how per-frame entropy values can be normalized and aggregated to obtain a confidence measure per unit and per word. We evaluate the proposed confidence measures on LibriSpeech test sets, and show that they are up to 2 and 4 times better than confidence estimation based on the maximum per-frame probability.
  arXiv  Detail & Related papers  (2022-12-16T20:27:40Z)
• Composably secure data processing for Gaussian-modulated continuous variable quantum key distribution [58.720142291102135]
  Continuous-variable quantum key distribution (QKD) employs the quadratures of a bosonic mode to establish a secret key between two remote parties. We consider a protocol with homodyne detection in the general setting of composable finite-size security. In particular, we analyze the high signal-to-noise regime which requires the use of high-rate (non-binary) low-density parity check codes.
  arXiv  Detail & Related papers  (2021-03-30T18:02:55Z)

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.