mlsanityData to 7 Oct 2026

Acoustic Progress Propagation for Long-Horizon Speculative Decoding in ASR

arXiv 2609.33245 ↗First posted 27 Sep 2026eess.AS cs.CLcited by 0

On the map: Efficient LLM inference and serving → Speculative decoding

Abstract (arXiv)

Speculative decoding accelerates autoregressive automatic speech recognition (ASR), but the acceptance length of alignment-aware drafters can saturate as the draft horizon increases. We propose a progress-aware speculative drafter that recurrently propagates an acoustic progress state across draft steps and feeds it back into audio cross-attention to guide token generation. We jointly train the drafter and progress predictor over variable draft horizons. On five ASR test sets, our method achieves lossless, macro-averaged end-to-end speedups of 1.657x and 1.227x over target-only autoregressive decoding with Qwen3-ASR-0.6B and Qwen3-ASR-1.7B, respectively. Relative to AnchorDraft, our method improves the macro-averaged speedup by 34.3% and 9.0%, respectively. Horizon sweeps show continued growth in acceptance length beyond the baselines' saturation. Code is available at https://github.com/yuanyuanjia71-spec/ProgDraft.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving. Builds on 16 of its 21 references that are in the map. First posted on arXiv 27 Sep 2026.

Builds on

16 of 21 references.

Closest ancestors · 6

Same direction, newest first

  1. 2026DFlash: Block Diffusion for Flash Speculative Decoding5 Feb
  2. 2025EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test3 Mar
  3. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty26 Jan
  4. 2024Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads19 Jan
  5. 2023Accelerating Large Language Model Decoding with Speculative Sampling2 Feb
  6. 2022Fast Inference from Transformers via Speculative Decoding30 Nov

Other references · 10

  1. 2017Attention Is All You Needcited by 59,932
  2. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 3,972
  3. 2022FLEURS: Few-shot Learning Evaluation of Universal Representations of Speechcited by 458
  4. 2023Scaling Speech Technology to 1,000+ Languagescited by 401
  5. 2021GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audiocited by 391
  6. 2018TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptationcited by 224
  7. 2026Qwen3-ASR Technical Reportcited by 105
  8. 2024Whisper in Medusa's Ear: Multi-head Efficient Decoding for Transformer-based ASRcited by 6
  9. 2025SpecASR: Accelerating LLM-based Automatic Speech Recognition via Speculative Decodingcited by 4
  10. 2026Alignment Drift in Single-Model Speculative Decoding for ASR: Mechanism, Correction, and Costcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Model-free Speculative Decoding for Transformer-based ASR with Token Map Drafting
  2. 2025Accelerating Autoregressive Speech Synthesis Inference With Speech Speculative Decoding
  3. 2026MDM-ASR: Bridging Accuracy and Efficiency in ASR with Diffusion-Based Non-Autoregressive Decoding ↗
  4. 2026Decoding Strategies for Diffusion-Based ASR: A Systematic Evaluation of Confidence-Based Thresholding ↗
  5. 2025Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
  6. 2025Principled Coarse-Grained Acceptance for Speculative Decoding in Speech
  7. 2026AS2D: Accelerating On-Demand Audio Understanding on Mobile Devices
  8. 2025Whisfusion: Parallel ASR Decoding with Masked Diffusion