Automatic speech recognition for under-resourced languages: A survey

By Laurent Besacier, Etienne Barnard, Alexey Karpov, Tanja Schultz. 2014

547 citations recorded by the source corpus.

Speech processing for under-resourced languages is an active field of research, which has experienced significant progress during the past decade. We propose, in this paper, a survey that focuses on automatic speech recognition (ASR) for these languages. The definition of under-resourced languages and the challenges associated to them are first defined. The main part of the paper is a literature review of the recent (last 8years) contributions made in ASR for under-resourced languages. Examples of past projects and future trends when dealing with under-resourced languages are also presented. We believe that this paper will be a good starting point for anyone interested to initiate research in (or operational development of) ASR for one or several under-resourced languages. It should be clear, however, that many of the issues and approaches presented here, apply to speech technology in general (text-to-speech synthesis for instance).

Subfields: Artificial Intelligence, Information Systems, Speech & Audio

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Speech & Audio

Citations in the corpus

Multilingual Speech Recognition with a Single End-to-End Model (2017, background): “However, in many cases the resources available to train large vocabulary continuous speech recognizers are severely limited [1].”
Unsupervised acoustic unit discovery for speech synthesis using discrete latent-variable neural networks (2019, background): “For many languages, it is difﬁcult or impossible to collect the annotated resources required for training supervised automatic speech recognition (ASR) models [1].”
An embedded segmental K-means model for unsupervised segmentation and clustering of speech (2017, neutral): “, unwritten or endangered languages [1].”
First automatic fongbe continuous speech recognition system: Development of acoustic models and language models (2016, background): “ASR applications have been successfully achieved for most western languages such as English, French, Italian etc., for Asian languages such as Chinese, Japanese, Indian etc, because of the large quantity and the availability of linguistic resources of these languages [2].”
Opportunities and Challenges for Artificial Intelligence in India (2018, background): “Like NLP, the field of ASR has also demarcated under-resourced languages as a special topic [8].”
Visually Grounded Learning of Keyword Prediction from Untranscribed Speech (2017, background): “for endangered or unwritten languages [6]; grounding speech using co-occurring visual contexts could be a way to train systems in such lowresource scenarios [7].”
An empirical evaluation of zero resource acoustic unit discovery (2017, unknown): “However, for a large majority of the world’s languages, it is often difﬁcult or even almost impossible to collect enough language resources to develop ASR systems with current standard ASR technology [1].”
Multilingual Sequence-to-Sequence Speech Recognition: Architecture, Transfer Learning, and Language Modeling (2018, methodology): “In this paper, we explore the multilingual training approaches [9, 10, 11] used in hybrid DNN/RNN-HMMs to incorporate them into the seq2seq models.”
Multimodal One-shot Learning of Speech and Images (2018, background): “This setting is relevant in modelling infant language acquisition, where models can be used to test particular cognitive hypotheses [4]; lowresource speech processing, where new concepts could be taught in an arbitrary language [5]; and robotics, where novel concepts must be acquired online from co-”
Language Independent End-to-End Architecture For Joint Language and Speech Recognition (result): “There have been a lot of prior studies on multilingual/languageindependent ASR (see [20] for more details).”
Neural Machine Translation for Low-resource Languages: A Survey (2021, background): “2An LRL is also known as under resourced, low-density, resource-poor, low data, or less-resourced language [16]”
Reliable detection of audio events in highly noisy environments (2015, background): “Audio analysis has been traditionally focused on the recognition of speech (Anusuya and Katti, 2010; Besacier et al., 2014), speaker identification (Cordella et al.”
Visually grounded cross-lingual keyword spotting in speech (2018, background): “For many languages, it is difficult or impossible to collect such annotated resources [1].”
Semantic Speech Retrieval With a Visually Grounded Model of Untranscribed Speech (2017, background): “However, current methods require large amounts of transcribed speech data, which are available only for a small fraction of the world’s languages (Besacier et al., 2014).”
Open-source Multi-speaker Speech Corpora for Building Gujarati, Kannada, Malayalam, Marathi, Tamil and Telugu Speech Synthesis Systems (2020, background): “Although the situation with the speech corpora availability for these languages has been improving, these languages are still considered by many to be low-resource (Besacier et al., 2014; Srivastava et al., 2018).”