Applied Speech and Text Processing for Low Resource Languages
Motivation and Scope of the Book
Out of over 7000 recognized languages worldwide, only a small proportion offer sufficient resources to build speech and natural language processing (NLP) technologies adequately. Developing such solutions for low-resource languages is challenging in multiple aspects. The dependency on deep learning over volumes of resources is quite conventional. Hence, the research in this domain is propelled by different ideas to train and validate systems, such as data augmentation, transfer learning, and hybrid multi-modal architectures, to name a few. This book aims at collecting such ideas, advent, and solutions for building speech/NLP technologies in low-resource scenarios.
Timeline
The submission guidelines and other detailed information can be found on the book website.
Full Chapter Abstract Submission: 20.02.2024 (Extended)
Abstract Acceptance/Rejection Notification: 25.02.2024
Full Chapter Submission: 15.03.2024
First Review Notification: 15.04.2024
Revised Version Notification: 15.05.2024
Final Acceptance/Rejection Notification: 30.05.2024
Camera Ready Chapter Submission: 10.06.2024
Table of Contents
We invite submissions of high-quality, original chapters addressing both theoretical and practical aspects, including their ethical and social implications. The Book aims to cover (but is not limited) the following topics:
Speech Processing for Low Resource Languages
Natural Language Processing for Low Resource
Deep learning methods for low-resource languages
End-to-end speech Recognition for Low-Resource Language
Development of a Speech Corpus for Low Resource Language
Multimodal architectures for social media analysis
Speech Recognition for Specific Dialects
Speech synthesis for low-resource language
Speech translation for low-resource language
Indigenous Language revitalization/preservation
Transfer learning applications
Leveraging Large pre-trained language models knowledge under few-shot and zero-shot in NLP tasks
Efficiently aligning acoustic and textual embeddings
Submission Guidelines
Authors should send their abstracts and chapters/articles through easy-chair only.
Abstract: The abstract must contain 300-400 words including title, chapter outline/content, and author (s) affiliation.
Chapter/article: Each chapter/article must contain 15-25 pages with approximately 9,500-12,500 words including references (contact the editors for additional page approval). Each chapter/article must be compliant with the Alt Text format. If you need more details, don't hesitate to contact the editors.
EDITORS
Publication
The edited book will be published by River Publisher.
CONTACT
All questions about submissions should be emailed to speech_nlp_book_river@googlegroups.com