Summary of Efficient Adaptation Of Multilingual Models For Japanese Asr, by Mark Bajo et al.
Efficient Adaptation of Multilingual Models for Japanese ASR
by Mark Bajo, Haruka Fukukawa, Ryuji Morita, Yuma Ogasawara
First submitted to arxiv on: 14 Dec 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Sound (cs.SD); Audio and Speech Processing (eess.AS)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The study explores the fine-tuning of OpenAI’s Whisper-Tiny, a multilingual Automatic Speech Recognition (ASR) model, to improve performance in Japanese. The authors compare Whisper-Tiny to monolingual models like ReazonSpeech and find that fine-tuning reduces the Character Error Rate (CER) from 32.7 to 14.7. LoRA and end-to-end training methods are used to adapt the model to Japanese-specific datasets. The results show that fine-tuning achieves strong language-specific performance while retaining flexibility, making it a scalable solution for ASR in resource-constrained environments and languages with complex writing systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study shows how to improve Automatic Speech Recognition (ASR) in Japanese by fine-tuning a multilingual model called Whisper-Tiny. The current best models are good at many languages but not perfect in one language. By making the model better for Japanese, we can get it to work more accurately. We tested different ways of fine-tuning and found that some methods worked much better than others. This is important because ASR technology can help people who are deaf or have difficulty hearing. |
Keywords
» Artificial intelligence » Cer » Fine tuning » Lora
