Experiments on Turkish ASR with Self-Supervised Speech Representation Learning

TL;DR

This paper explores Turkish automatic speech recognition using self-supervised learning with HUBERT, trained on 6,500 hours of YouTube data, highlighting current limitations in robustness for real-world applications.

Contribution

It presents the first large-scale pre-training of HUBERT for Turkish ASR using extensive online data, analyzing its performance and limitations.

Findings

Models lack robustness against real-world disturbances

Significant errors identified in accent and noise variations

Pre-training improves baseline but needs further enhancement

Abstract

While the Turkish language is listed among low-resource languages, literature on Turkish automatic speech recognition (ASR) is relatively old. In this report, we present our findings on Turkish ASR with speech representation learning using HUBERT. We investigate pre-training HUBERT for Turkish with large-scale data curated from online resources. We pre-train our model using 6,500 hours of speech data from YouTube. The results show that the models are not ready for commercial use since they are not robust against disturbances that typically occur in real-world settings such as variations in accents, slang, background noise and interference. We analyze typical errors and the limitations of the models for use in commercial settings.