Abstract:This paper investigates efficient methods for utilizing text-only data to improve speech recognition, focusing on encoder-dominated models that facilitate faster recognition. We provide a comprehensive comparison of techniques to integrate text-only data, including modality matching and dynamic downsampling to reach text-level representations within the encoder. Our experiments on the LibriSpeech corpus show that a larger encoder with a smaller decoder can equal or surpass the performance of architectures with larger decoders. We demonstrate that simple configurations, such as random duration models, are often more effective than complex alternatives, significantly simplifying the training pipeline. All code and recipes are made publicly available.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE) |
| Cite as: | arXiv:2604.26514 [cs.CL] |
| (or arXiv:2604.26514v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2604.26514 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Albert Zeyer [view email]
[v1]
Wed, 29 Apr 2026 10:28:00 UTC (100 KB)