Multi-Level Embedding Conformer Framework for Bengali Automatic Speech Recognition

TL;DR

This paper introduces a multi-level embedding Conformer framework for Bengali ASR, integrating phoneme, syllable, and wordpiece information to improve recognition accuracy in a low-resource language.

Contribution

It proposes a novel multi-level embedding fusion mechanism within a Conformer-CTC model specifically designed for Bengali ASR, enhancing phonetic and contextual feature capture.

Findings

Achieved a WER of 10.01% on Bengali speech data.

Demonstrated the effectiveness of multi-granular linguistic embeddings.

Showed improved recognition performance over baseline models.

Abstract

Bengali, spoken by over 300 million people, is a morphologically rich and lowresource language, posing challenges for automatic speech recognition (ASR). This research presents an end-to-end framework for Bengali ASR, building on a Conformer-CTC backbone with a multi-level embedding fusion mechanism that incorporates phoneme, syllable, and wordpiece representations. By enriching acoustic features with these linguistic embeddings, the model captures fine-grained phonetic cues and higher-level contextual patterns. The architecture employs early and late Conformer stages, with preprocessing steps including silence trimming, resampling, Log-Mel spectrogram extraction, and SpecAugment augmentation. The experimental results demonstrate the strong potential of the model, achieving a word error rate (WER) of 10.01% and a character error rate (CER) of 5.03%. These results demonstrate the effectiveness of combining multi-granular linguistic information with acoustic modeling, providing a scalable approach for low-resource ASR development.