On the difficulty of a distributional semantics of spoken language

TL;DR

This paper explores the challenges of developing distributional semantic models for spoken language, highlighting the difficulties and potential approaches to abstract from surface variability in speech to learn meaningful representations.

Contribution

It investigates the adaptation of unsupervised semantic learning methods from written to spoken language and evaluates simple models on synthetic and human speech datasets.

Findings

Models learn some semantic representations from synthetic speech

Results on human speech are inconclusive

Discusses inherent challenges in natural spoken language semantics

Abstract

In the domain of unsupervised learning most work on speech has focused on discovering low-level constructs such as phoneme inventories or word-like units. In contrast, for written language, where there is a large body of work on unsupervised induction of semantic representations of words, whole sentences and longer texts. In this study we examine the challenges of adapting these approaches from written to spoken language. We conjecture that unsupervised learning of the semantics of spoken language becomes feasible if we abstract from the surface variability. We simulate this setting with a dataset of utterances spoken by a realistic but uniform synthetic voice. We evaluate two simple unsupervised models which, to varying degrees of success, learn semantic representations of speech fragments. Finally we present inconclusive results on human speech, and discuss the challenges inherent in learning distributional semantic representations on unrestricted natural spoken language.