Summary of Speech Prefix-tuning with Rnnt Loss For Improving Llm Predictions, by Murali Karthick Baskar et al.
Speech Prefix-Tuning with RNNT Loss for Improving LLM Predictions
by Murali Karthick Baskar, Andrew Rosenberg, Bhuvana Ramabhadran, Neeraj Gaur, Zhong Meng
First submitted to arxiv on: 20 Jun 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); 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 This paper proposes novel techniques for applying Large Language Models (LLMs) to Automatic Speech Recognition (ASR), addressing constraints in recent works that utilize prefixLM-type models. The authors introduce speech prefix-tuning, which optimizes speech prefixes for better ASR performance, and language-based soft prompting to further improve results with frozen LLMs. Experimental analysis on real-time testsets from 10 Indic languages demonstrates the effectiveness of these approaches, yielding a 12% relative improvement in Word Error Rate (WER) over baseline fine-tuned LLMs. The proposed methods also lead to a 31% relative improvement over basic soft-prompting prefixLM when using frozen LLMs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps make it easier for computers to recognize spoken language. It presents new ways to use large language models in automatic speech recognition, which is important for applications like voice assistants and speech-to-text systems. The authors tested their ideas on 10 Indian languages and found that they work well, even when using pre-trained models that aren’t specifically designed for speech recognition. |
Keywords
» Artificial intelligence » Prompting
