We introduce a new collection of spoken English audio suitable for training speech recognition systems under limited or no supervision. It is derived from open-source audio books from the LibriVox project. It contains over 60K hours of audio, which is, to our knowledge, the largest freely-available corpus of speech. The audio has been segmented using voice activity detection and is tagged with SNR, speaker ID and genre descriptions. Additionally, we provide baseline systems and evaluation metrics working under three settings: (1) the zero resource/unsupervised setting (ABX), (2) the semisupervised setting (PER, CER) and (3) the distant supervision setting (WER). Settings (2) and (3) use limited textual resources (10 minutes to 10 hours) aligned with the speech. Setting (3) uses large amounts of unaligned text. They are evaluated on the standard LibriSpeech dev and test sets for comparison with the supervised state-of-the-art.Index Terms-unsupervised and semi-supervised learning, distant supervision, dataset, zero-and low resource ASR.

We describe a new challenge aimed at discovering subword and word units from raw speech. This challenge is the followup to the Zero Resource Speech Challenge 2015. It aims at constructing systems that generalize across languages and adapt to new speakers. The design features and evaluation metrics of the challenge are presented and the results of seventeen models are discussed.Index Terms-zero resource speech technology, subword modeling, acoustic unit discovery, unsupervised term discovery

No abstract

Fixed-length embeddings of words are very useful for a variety of tasks in speech and language processing. Here we systematically explore two methods of computing fixed-length embeddings for variable-length sequences. We evaluate their susceptibility to phonetic and speaker-specific variability on English, a high resource language, and Xitsonga, a low resource language, using two evaluation metrics: ABX word discrimination and ROC-AUC on same-different phoneme n-grams. We show that a simple downsampling method supplemented with length information can be competitive with the variable-length input feature representation on both evaluations. Recurrent autoencoders trained without supervision can yield even better results at the expense of increased computational complexity.

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.

We present "DiViMe", an open-source virtual machine aimed at packaging speech technology for real-life data, and developed in the context of the "Analyzing Children's Language Environments across the World" Project. This first release focuses on Speech Activity Detection, Speaker Diarization, and their evaluation. The present paper introduces the set of included tools and the current workflow, which is focused on making minimal assumptions regarding users' technical skills. Additionally, we show how the current DiViMe tools fare against three sets of challenging data. In a first experiment, we look at performance with samples extracted from daylong recordings gathered using the LENA TM system from English-learning children. We find that the performance of the tools currently in DiViMe is not far from that achieved by the LENA TM proprietary software. In a second experiment, we generalize to other samples of child-centered daylong files, gathered with non-LENA TM hardware from non-English-learning children, showing that performance does not degrade in this condition. Finally, we report on performance in the DiHARD 2018 Challenge Test Data. Originally conceived in the "Speech Recognition Virtual Kitchen", DiViMe is a promising platform for packaging speech technology tools for widespread re-use, with potential impact on both fundamental and applied speech and language research.

We present the Zero Resource Speech Challenge 2019, which proposes to build a speech synthesizer without any text or phonetic labels: hence, TTS without T (text-to-speech without text). We provide raw audio for a target voice in an unknown language (the Voice dataset), but no alignment, text or labels. Participants must discover subword units in an unsupervised way (using the Unit Discovery dataset) and align them to the voice recordings in a way that works best for the purpose of synthesizing novel utterances from novel speakers, similar to the target speaker's voice. We describe the metrics used for evaluation, a baseline system consisting of unsupervised subword unit discovery plus a standard TTS system, and a topline TTS using gold phoneme transcriptions. We present an overview of the 19 submitted systems from 10 teams and discuss the main results.

Recent studies have investigated siamese network architectures for learning invariant speech representations using samedifferent side information at the word level. Here we investigate systematically an often ignored component of siamese networks: the sampling procedure (how pairs of same vs. different tokens are selected). We show that sampling strategies taking into account Zipf's Law, the distribution of speakers and the proportions of same and different pairs of words significantly impact the performance of the network. In particular, we show that word frequency compression improves learning across a large range of variations in number of training pairs. This effect does not apply to the same extent to the fully unsupervised setting, where the pairs of same-different words are obtained by spoken term discovery. We apply these results to pairs of words discovered using an unsupervised algorithm and show an improvement on state-of-the-art in unsupervised representation learning using siamese networks.