EC Dept. Institute of Technology Nirma University hosted an insightful lecture on “Recent Advances in Speech Technologies for Indian Languages” by Prof. Srinivasan Umesh from the Electrical Engineering department, IIT Madras.
Key Takeaways:
– Overview of exciting work in Speech Technologies for Indian Languages at SPRING lab IIT-Madras.
– Insights into three broad architectures for ASR: encoder-decoder, CTC, and transducer-based approaches
– State-of-the-art multilingual TTS and open-source models and data
– Demo of ASR, TTS, and speech-to-speech translation systems
– Introduction to speech foundation models: ccc-wav2vec2.0 and data2vec-aqc
– Future directions in building Speech-LLMs
Thanks to Prof. Umesh for sharing his expertise and knowledge with us! His talk highlighted the progress and potential of speech technologies for Indian languages.
