Papers/2608.20346
🧪 Test?View on arXiv

Building and Evaluating a Synthetic Bengali Speech Resource for Telecom Customer Care

speech synthesisASRdataset creationcustomer care
2608.20346
Builder Relevance
80%
2h ago

Abstract

The paper presents a synthetic Bengali speech dataset for telecom customer-care scenarios.

Reality Card

Core Claim

A synthetic Bengali speech dataset containing 10,000 audio-text pairs was created and evaluated, demonstrating strong text-audio consistency.

Method / Result

Average WER of 2.54% and average CER of 0.59% across 10,000 samples.

Limitations

The paper discusses limitations of synthetic speech and STT-based evaluation.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers