BanglaRobustNet: A Hybrid Denoising-Attention Architecture for Robust Bangla Speech Recognition
Md Sazzadul Islam Ridoy, Mubaswira Ibnat Zidney, Sumi Akter and Md. Aminur Rahman · 2026
Read BanglaRobustNet: A Hybrid Denoising-Attention Architecture for Robust Bangla Speech Recognition by Md Sazzadul Islam Ridoy, Mubaswira Ibnat Zidney, Sumi Akter and Md. Aminur Rahman (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-17679.txt.
Description
Md Sazzadul Islam Ridoy and three co-authors present BanglaRobustNet for Bangla speech recognition. Bangla is one of the world's most spoken languages. It stays underrepresented in top speech research. Noise and speaker diversity make the gap worse. Their hybrid framework builds on Wav2Vec-BERT. It pairs denoising with attention in one design.
Robust recognition serves Bangla speakers in real settings. Streets, markets, and homes are noisy by nature. A model trained for quiet rooms fails there. Denoising up front gives the recognizer cleaner input. Attention then weighs the sounds that matter. The work pushes low-resource language tech forward. It shows gains need not wait for huge datasets.
Catalog record
| CREATOR | Md Sazzadul Islam Ridoy, Mubaswira Ibnat Zidney, Sumi Akter and Md. Aminur Rahman |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-17679