Shubham Korde
← Back to the 12-chapter story

IIT Kanpur

Research, speech and healthcare

A speech dataset and pilot for intrapartum care.

Hindi speechSpeech datasetCONCEPTUAL DIAGRAM · NCC 2023
Hindi speech recordings collected for a speech-recognition dataset.

During college, I worked on speech processing and an AI-for-social-good project with CARE India and Google Research India. It focused on intrapartum care, meaning care during labour and childbirth, in settings where experienced nurses may not always be available.

I led the project and worked on an AI-based solution. We initiated a pilot at four medical facilities in Bihar. Any wider expansion depended on the trials; it was proposed, not completed.

I also co-authored “Parturition Hindi Speech Dataset for Automatic Speech Recognition,” presented at NCC 2023. The paper was about a Hindi speech dataset for automatic speech recognition. The pilot was part of the broader care project. They were related efforts, with different outputs.

The project brought speech research into a specific healthcare setting. People worked on different parts of it, from research to the pilot.

A paper and a pilot

I co-authored the paper presented at NCC 2023 and led the intrapartum-care project. The project initiated a pilot at four facilities in Bihar.

The research and the setting

The planned expansion depended on the trial results. The paper documented a Hindi speech dataset; the pilot belonged to the broader care project.