arXiv cs.CL· Simon Berger, Larissa Kleppel, Ralf Schl\"uter·· 13 小时前AI 评分22
CTC 与神经 transducer 端到端 ASR 模型的搜索空间复杂度与效率系统对比
A Systematic Comparison of Search Space Complexity and Efficiency for Time-Synchronous End-to-End ASR Models
AI 导读
研究系统对比了 CTC 与神经 transducer 端到端 ASR 模型在不同剪枝策略和 LM 配置下的搜索空间复杂度与效率。这类模型多数时候只需很小的搜索空间，但在不确定时段会出现假设数量的临时高峰；基于分数的剪枝能随模型不确定性动态调整搜索，显著降低实时率（RTF）。结果说明低固定 beam 剪枝存在局限，并为端到端 ASR 的高效搜索提供了实践指南。
正文 · 原文
Abstract:Search is a central component of Automatic Speech Recognition (ASR) systems and has been deeply investigated for classical ASR modeling based on Hidden Markov Models (HMMs). However, in work on current sequence-to-sequence ASR architectures it is often simplified to a small fixed-size beam search without external Language Model (LM). In this work, we present a comparative analysis of search space complexity and efficiency for Connectionist Temporal Classification (CTC) and neural transducer models under different pruning strategies and LM configurations. We show that these end-to-end (E2E) models work with a small search space most of the time but exhibit large temporary peaks in hypothesis count during periods of uncertainty. Score-based pruning dynamically adapts the search to model uncertainty, leading to a significantly lower Real-Time Factor (RTF). Our results highlight the limitations of search with low fixed-size beam pruning and provide practical guidelines for efficient search in E2E ASR.
| Comments: | Accepted at IEEE SLT 2026. 7 pages, 4 figures |
| Subjects: | Computation and Language (cs.CL) |
| ACM classes: | I.2.7 |
| Cite as: | arXiv:2610.31259 [cs.CL] |
| (or arXiv:2610.31259v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31259 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Simon Berger [view email]
[v1]
Thu, 22 Oct 2026 13:35:50 UTC (137 KB)
来源：arXiv cs.CL · arxiv.org