Perceptimatic: A human speech perception benchmark for unsupervised subword modelling

TL;DR

This paper introduces Perceptimatic, a benchmark dataset for comparing human speech perception with models on phone discrimination tasks, highlighting differences between model and human perceptual spaces.

Contribution

It provides a new open dataset and a method to compare human and model speech perception, applied to existing models from the Zero Resource Speech Challenge.

Findings

Supervised monolingual HMM-GMM models differ from human perceptual space.

Unsupervised and multilingual models show different perceptual representations.

The dataset enables detailed comparison of human and model speech perception.

Abstract

In this paper, we present a data set and methods to compare speech processing models and human behaviour on a phone discrimination task. We provide Perceptimatic, an open data set which consists of French and English speech stimuli, as well as the results of 91 English- and 93 French-speaking listeners. The stimuli test a wide range of French and English contrasts, and are extracted directly from corpora of natural running read speech, used for the 2017 Zero Resource Speech Challenge. We provide a method to compare humans' perceptual space with models' representational space, and we apply it to models previously submitted to the Challenge. We show that, unlike unsupervised models and supervised multilingual models, a standard supervised monolingual HMM-GMM phone recognition system, while good at discriminating phones, yields a representational space very different from that of human native listeners.