mlsanityData to 14 Oct 2026

A Generalized Formalism of Auto-Regressive Decoding for Speech Processing

arXiv 2606.20714 ↗First posted 16 Jun 2026cs.SD cs.LG eess.AScited by 0

Not on the map. Why?

Abstract (arXiv)

In speech processing, most state-of-the-art sequence prediction models rely on auto-regressive (AR) strategies to generate output sequences based on the raw predictions of the model. Despite their crucial role in the inference process, a comprehensive overview of AR strategies as a unified field is lacking, due largely to implicit and multiple definitions of next-token decoding. This context complicates the choice, comparison, and evaluation of strategies, while creating inconsistencies in the characterization of approaches as auto-regressive or not. We begin by setting explicit inclusion criteria for the field of AR search in speech processing, and derive a generalized theoretical framework to categorize and report on search strategies for neural models. We show the capabilities of this formalism in simplifying the design of benchmarks centered around the decoding process, allowing for ablation studies that are focused on search strategies.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 32 of its 45 references that are in the map. First posted on arXiv 16 Jun 2026.

Builds on

32 of 45 references.

References · 32

  1. 2019The Curious Case of Neural Text Degenerationcited by 2,916
  2. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  3. 2023Faith and Fate: Limits of Transformers on Compositionalitycited by 503
  4. 2017Non-Autoregressive Neural Machine Translationcited by 463
  5. 2018Blockwise Parallel Decoding for Deep Autoregressive Modelscited by 323
  6. 2017Beam Search Strategies for Neural Machine Translationcited by 232
  7. 2024The pitfalls of next-token predictioncited by 160
  8. 2020Is MAP Decoding All You Need? The Inadequacy of the Mode in Neural Machine Translationcited by 138
  9. 2024From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Modelscited by 137
  10. 2023End-to-End Speech Recognition: A Surveycited by 129
  11. 2020On Exposure Bias, Hallucination and Domain Shift in Neural Machine Translationcited by 127
  12. 2023Accelerating Transformer Inference for Translation via Parallel Decodingcited by 124
20 more
  1. 2022Truncation Sampling as Language Model Desmoothingcited by 123
  2. 2024VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wildcited by 120
  3. 2018Learning to Write with Cooperative Discriminatorscited by 119
  4. 2024A Thorough Examination of Decoding Methods in the Era of LLMscited by 114
  5. 2022Locally Typical Samplingcited by 111
  6. 2022Why Exposure Bias Matters: An Imitation Learning Perspective of Error Accumulation in Language Generationcited by 108
  7. 2022On Decoding Strategies for Neural Text Generatorscited by 90
  8. 2019Comparison of Diverse Decoding Methods from Conditional Language Modelscited by 78
  9. 2022A Survey on Non-Autoregressive Generation for Neural Machine Translation and Beyondcited by 64
  10. 2021Machine Translation Decoding beyond Beam Searchcited by 55
  11. 2021Instantaneous Grammatical Error Correction with Shallow Aggressive Decodingcited by 53
  12. 2022Grammatical Error Correction: A Survey of the State of the Artcited by 43
  13. 2020Best-First Beam Searchcited by 40
  14. 2024Towards Controllable Speech Synthesis in the Era of Large Language Models: A Systematic Surveycited by 23
  15. 2020Decoding Methods for Neural Narrative Generationcited by 18
  16. 2025Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesiscited by 10
  17. 2023CTC-based Non-autoregressive Speech Translationcited by 9
  18. 2023End-to-End Speech-to-Text Translation: A Surveycited by 9
  19. 2022Language Model Decoding as Likelihood-Utility Alignmentcited by 5
  20. 2024Enabling Beam Search for Language Model-Based Text-to-Speech Synthesiscited by 2

Built on by

Not cited yet.