Summary of Multistage Fine-tuning Strategies For Automatic Speech Recognition in Low-resource Languages, by Leena G Pillai et al.
Multistage Fine-tuning Strategies for Automatic Speech Recognition in Low-resource Languages
by Leena G Pillai, Kavya Manohar, Basil K Raju, Elizabeth Sherly
First submitted to arxiv on: 7 Nov 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 proposed novel multistage fine-tuning strategy aims to enhance automatic speech recognition (ASR) performance in low-resource languages using OpenAI’s Whisper model. The approach builds ASR models for languages with limited digital resources by sequentially adapting the model across linguistically similar languages. This is demonstrated on the Malasar language, a Dravidian language spoken by approximately ten thousand people in the Western Ghats of South India. The paper presents an intermediate Tamil ASR model fine-tuned on Malasar data, achieving a word error rate (WER) of 51.9%. Further WER reduction to 47.3% is achieved through punctuation removal in post-processing, highlighting the effectiveness of sequential multistage fine-tuning and targeted post-processing for ASR system development in low-resource languages. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps build a speech recognition model for people who speak Malasar, a language with limited digital resources. The researchers used OpenAI’s Whisper model to develop an ASR model that can understand spoken words in the Malasar language. They fine-tuned the model by first creating an intermediate model for Tamil, a closely related language with more available data, and then adapting this model for Malasar. This approach achieved better results than trying to build the model directly from Malasar data alone. |
Keywords
» Artificial intelligence » Fine tuning
