Papers/2609.20828
🧪 Test?View on arXiv

Beyond WER: Entity and Disfluency Recall in Accented Conversational ASR

Author1, Author2, Author3, Author4, Author5

ASRentity recognitionfine-tuningaccented speech
2609.20828
Builder Relevance
80%
1h ago

Abstract

This paper presents a novel pipeline for improving entity and disfluency recall in accented conversational ASR systems.

Reality Card

Core Claim

The proposed pipeline achieves 80-85% entity recall and 76-86% filler recall, significantly improving upon previous benchmarks.

Method / Result

Entity recall improved from 53-55% to 80-85% and filler recall from <5% to 76-86%.

Limitations

The study's results may be limited to the specific accents and regions tested, potentially affecting generalizability.

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