An auto-encoder based approach to unsupervised learning of subword units

摘要

In this paper we propose an autoencoder-based method for the unsupervised identification of subword units. We experiment with different types and architectures of autoencoders to asses what autoencoder properties are most important for this task. We first show that the encoded representation of speech produced by standard autencoders is more effective than Gaussian posteriorgrams in a spoken query classification task. Finally we evaluate the subword inventories produced by the proposed method both in terms of classification accuracy in a word classification task (with lexicon size up to 263 words) and in terms of consistency between subword transcription of different word examples of a same word type. The evaluation is carried out on Italian and American English datasets.