QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus

TL;DR

QASR is the largest annotated Arabic speech corpus from broadcast media, enabling advancements in speech recognition, dialect identification, and NLP tasks with extensive transcribed speech and auxiliary language data.

Contribution

This paper introduces QASR, the largest multi-dialect Arabic speech dataset with detailed annotations, and provides baseline results for speech recognition and NLP tasks.

Findings

End-to-end speech recognition achieves competitive WER.

Baseline results for Arabic punctuation restoration.

Resource availability for future research.

Abstract

We introduce the largest transcribed Arabic speech corpus, QASR, collected from the broadcast domain. This multi-dialect speech dataset contains 2,000 hours of speech sampled at 16kHz crawled from Aljazeera news channel. The dataset is released with lightly supervised transcriptions, aligned with the audio segments. Unlike previous datasets, QASR contains linguistically motivated segmentation, punctuation, speaker information among others. QASR is suitable for training and evaluating speech recognition systems, acoustics- and/or linguistics- based Arabic dialect identification, punctuation restoration, speaker identification, speaker linking, and potentially other NLP modules for spoken data. In addition to QASR transcription, we release a dataset of 130M words to aid in designing and training a better language model. We show that end-to-end automatic speech recognition trained on QASR reports a competitive word error rate compared to the previous MGB-2 corpus. We report baseline results for downstream natural language processing tasks such as named entity recognition using speech transcript. We also report the first baseline for Arabic punctuation restoration. We make the corpus available for the research community.