Unsupervised Acoustic Unit Discovery by Leveraging a Language-Independent Subword Discriminative Feature Representation

Feng, Siyuan; Żelasko, Piotr; Moro-Velázquez, Laureano; Scharenborg, Odette

doi:10.21437/interspeech.2021-1664

Cited by 6 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While APC is not a model of infant learning per se, it is one of the simplest and hence perhaps the most easy to understand of modern machine learning methods for unsupervised learning. The latent representations APC learns have been shown to be applicable to a range of speech tasks, including phone classification (Chung et al, 2019) and subword modeling (Feng, Żelasko, Moro-Velázquez, & Scharenborg, 2021; see also Yang et al (2021)). Central to modeling purposes, APC learning criterion aligns with the idea of predictive processing in human perception (e.g., Friston, 2010;Rao & Ballard, 1999;Babineau, Havron, Dautriche, de Carvalho, & Christophe, 2022) and a similar learning mechanism can thereby be assumed to be available to any mammalian learner.…”

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Introducing Meta‐analysis in the Evaluation of Computational Models of Infant Language Development

2023

View full text Add to dashboard Cite

Computational models of child language development can help us understand the cognitive underpinnings of the language learning process, which occurs along several linguistic levels at once (e.g., prosodic and phonological). However, in light of the replication crisis, modelers face the challenge of selecting representative and consolidated infant data. Thus, it is desirable to have evaluation methodologies that could account for robust empirical reference data, across multiple infant capabilities. Moreover, there is a need for practices that can compare developmental trajectories of infants to those of models as a function of language experience and development. The present study aims to take concrete steps to address these needs by introducing the concept of comparing models with large‐scale cumulative empirical data from infants, as quantified by meta‐analyses conducted across a large number of individual behavioral studies. We formalize the connection between measurable model and human behavior, and then present a conceptual framework for meta‐analytic evaluation of computational models. We exemplify the meta‐analytic model evaluation approach with two modeling experiments on infant‐directed speech preference and native/non‐native vowel discrimination.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Introducing Meta‐analysis in the Evaluation of Computational Models of Infant Language Development

2023

View full text Add to dashboard Cite

show abstract

“…Finally, several recent studies have attempted a combination of Kempton's approach with the unsupervised clustering approach: cross-lingual ASR is used to annotate the articulatory features of an unknown language, which are then clustered to form unsupervised phonelike units [56,57]. To our knowledge, only two of these papers [71,72] directly evaluated phone inventory NMI or F1; using oracle cluster combination strategies that are standard in the field of unsupervised phone discovery, [72] achieved F1=64.14% for cross-lingual automatic phone inventory estimation.…”

Section: Related Workmentioning

confidence: 99%

Discovering Phonetic Inventories with Crosslingual Automatic Speech Recognition

Żelasko¹,

Feng²,

Moro-Velázquez³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

The high cost of data acquisition makes Automatic Speech Recognition (ASR) model training problematic for most existing languages, including languages that do not even have a written script, or for which the phone inventories remain unknown. Past works explored multilingual training, transfer learning, as well as zero-shot learning in order to build ASR systems for these low-resource languages. While it has been shown that the pooling of resources from multiple languages is helpful, we have not yet seen a successful application of an ASR model to a language unseen during training. A crucial step in the adaptation of ASR from seen to unseen languages is the creation of the phone inventory of the unseen language.The ultimate goal of our work is to build the phone inventory of a language unseen during training in an unsupervised way without any knowledge about the language. In this paper, we 1) investigate the influence of different factors (i.e., model architecture, phonotactic model, type of speech representation) on phone recognition in an unknown language; 2) provide an analysis of which phones transfer well across languages and which do not in order to understand the limitations of and areas for further improvement for automatic phone inventory creation; and 3) present different methods to build a phone inventory of an unseen language in an unsupervised way. To that

show abstract

“…In the last decade, unsupervised learning of acoustic models for ASR has gained increasing research interests [2]- [4]. It aims at discovering [2], [5], [6] (also referred to as acoustic unit discovery; AUD) or modeling [3], [4], [7] (also referred to as unsupervised subword modeling; USM) a set of basic speech units that represents all the sounds in the language in a zero-resource scenario, i.e., with only untranscribed data available. This research field aims to pave the way to developing high-performance ASR systems for languages that have very limited or no transcribed data.…”

Section: Introductionmentioning

confidence: 99%

The effectiveness of self-supervised representation learning in zero-resource subword modeling

Feng

Scharenborg

2021

2021 55th Asilomar Conference on Signals, Systems, and Computers

Self Cite

View full text Add to dashboard Cite

For a language with no transcribed speech available (the zero-resource scenario), conventional acoustic modeling algorithms are not applicable. Recently, zero-resource acoustic modeling has gained much interest. One research problem is unsupervised subword modeling (USM), i.e., learning a feature representation that can distinguish subword units and is robust to speaker variation. Previous studies showed that self-supervised learning (SSL) has the potential to separate speaker and phonetic information in speech in an unsupervised manner, which is highly desired in USM. This paper compares two representative SSL algorithms, namely, contrastive predictive coding (CPC) and autoregressive predictive coding (APC), as a front-end method of a recently proposed, state-of-the art two-stage approach, to learn a representation as input to a back-end cross-lingual DNN. Experiments show that the bottleneck features extracted by the back-end achieved state of the art in a subword ABX task on the Libri-light and ZeroSpeech databases. In general, CPC is more effective than APC as the front-end in our approach, which is independent of the choice of the out-domain language identity in the back-end cross-lingual DNN and the training data amount. With very limited training data, APC is found similar or more effective than CPC when test data consists of long utterances.

show abstract

Unsupervised Acoustic Unit Discovery by Leveraging a Language-Independent Subword Discriminative Feature Representation

Cited by 6 publications

References 27 publications

Introducing Meta‐analysis in the Evaluation of Computational Models of Infant Language Development

Introducing Meta‐analysis in the Evaluation of Computational Models of Infant Language Development

Discovering Phonetic Inventories with Crosslingual Automatic Speech Recognition

The effectiveness of self-supervised representation learning in zero-resource subword modeling

Contact Info

Product

Resources

About