Completely Unsupervised Phoneme Recognition by Adversarially Learning Mapping Relationships from Audio Embeddings
arXiv:1804.00316
Abstract
Unsupervised discovery of acoustic tokens from audio corpora without annotation and learning vector representations for these tokens have been widely studied. Although these techniques have been shown successful in some applications such as query-by-example Spoken Term Detection (STD), the lack of mapping relationships between these discovered tokens and real phonemes have limited the down-stream applications. This paper represents probably the first attempt towards the goal of completely unsupervised phoneme recognition, or mapping audio signals to phoneme sequences without phoneme-labeled audio data. The basic idea is to cluster the embedded acoustic tokens and learn the mapping between the cluster sequences and the unknown phoneme sequences with a Generative Adversarial Network (GAN). An unsupervised phoneme recognition accuracy of 36% was achieved in the preliminary experiments.
References in corpus (6)
- Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
- Categorical Reparameterization with Gumbel-Softmax
- The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
- A segmental framework for fully-unsupervised large-vocabulary speech recognition
- Segmental Audio Word2Vec: Representing Utterances as Sequences of Vectors with Applications in Spoken Term Detection
- Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data