Dear all,
Four positions are available for projects on audio-based archival content retrieval and speech recognition at MADHAV lab, IIT Kanpur:
These projects have been sponsored by Prasar Bharati, India’s Public Service Broadcaster, and involve working closely with Prasar Bharati on their data.
One project focuses on audio-based content retrieval, where the query is an audio or text, and the search content is audio. We will be developing state-of-the-art technologies using deep learning methods for various tasks such as
- Music Information Retrieval
- Audio Search
- Human in the loop learning
And the second project focuses on building automatic speech recognition (ASR) system for English, Hindi, and some other Indian languages for generating subtitles. It is a large vocabulary continuous speech recognition (LVCSR) task.
Thanks,
Vipul Arora