The Zero Resource Speech Challenge 2020: Discovering discrete subword and word units
This work addresses the challenge of unsupervised speech representation learning for researchers in speech processing, but it is incremental as it builds on previous benchmarks.
The Zero Resource Speech Challenge 2020 tackled the problem of learning speech representations from raw audio without labels by focusing on discovering subword and word units, with results from twenty models showing progress in unsupervised speech learning.
We present the Zero Resource Speech Challenge 2020, which aims at learning speech representations from raw audio signals without any labels. It combines the data sets and metrics from two previous benchmarks (2017 and 2019) and features two tasks which tap into two levels of speech representation. The first task is to discover low bit-rate subword representations that optimize the quality of speech synthesis; the second one is to discover word-like units from unsegmented raw speech. We present the results of the twenty submitted models and discuss the implications of the main findings for unsupervised speech learning.