OOD-Speech: A Large Bengali Speech Recognition Dataset for Out-of-Distribution Benchmarking

Published in Interspeech, 2023

A benchmark dataset for robust Bengali speech recognition across distribution shifts and acoustic variation.

Recommended citation: Dip, S. S., Alam, S., Tasnim, N., …, Sushmit, A. (2023). "OOD-Speech: A Large Bengali Speech Recognition Dataset for Out-of-Distribution Benchmarking." Interspeech.
Download Paper