Unsupervised Speech Recognition Via Segmental Empirical Output Distribution Matching

Abstract

We consider the problem of training speech recognition systems without using any labeled data, under the assumption that the learner can only access to the input utterances and a phoneme language model estimated from a non-overlapping corpus. We propose a fully unsupervised learning algorithm that alternates between solving two sub-problems: (i) learn a phoneme classifier for a given set of phoneme segmentation boundaries, and (ii) refining the phoneme boundaries based on a given classifier. To solve the first sub-problem, we introduce a novel unsupervised cost function named Segmental Empirical Output Distribution Matching, which generalizes the work in (Liu et al., 2017) to segmental structures. For the second sub-problem, we develop an approximate MAP approach to refining the boundaries obtained from Wang et al. (2017). Experimental results on TIMIT dataset demonstrate the success of this fully unsupervised phoneme recognition system, which achieves a phone error rate (PER) of 41.6%

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyyeh2018unsupervised

Related papers

• Towards Unsupervised Speech Recognition Without Pronunciation Models (2024)0.00
• Self-supervised Contrastive Learning For Unsupervised Phoneme Segmentation (2020)12.68
• Blind Phoneme Segmentation With Temporal Prediction Errors (2016)8.35
• Unsupervised Speech Recognition (2021)0.00
• Unsupervised Speech Segmentation: A General Approach Using Speech Language Models (2025)2.60
• Completely Unsupervised Phoneme Recognition By Adversarially Learning Mapping Relationships From Audio Embeddings (2018)0.00
• Unsupervised Speech Segmentation And Variable Rate Representation Learning Using Segmental Contrastive Predictive Coding (2021)9.92
• REBORN: Reinforcement-learned Boundary Segmentation With Iterative Training For Unsupervised ASR (2024)2.26