The Paris Graduate School of Linguistics (PGSL) is a newly-formed Paris-area graduate program covering all areas of language science.

It offers a comprehensive Master curriculum integrating advanced study and research, in close connection with PhD programs as well as with the Empirical Foundations of Linguistics consortium. 

Research plays a central part in the program, and students also take elective courses to develop an interdisciplinary outlook. Prior knowledge of French is not required.

For more details, please see https://paris-gsl.org/index.html

New funding opportunity: https://u-paris.fr/en/call-for-applications-international-students-miem-scholarship-program/

 Application deadline : February 1st 2021 (program starting September 1st 2021)

PGSL is funded by Smarts-UP (Student-centered iMproved, Active Research-based Training Strategy at Université de Paris) through the ANR SFRI grant « Grandes universités de recherche » (PIA3) 2020-2029.

Top

Cambridge's Machine Learning and Machine Intelligence MPhil

Are you interested in speech and language processing, computer vision & robotics, human-computer interaction, or machine learning? Please consider applying to the University of Cambridge?s Machine Learning and Machine Intelligence (MLMI) MPhil programme. 

The MLMI MPhil is an elite 11 month programme with a small cohort of about 30 students each year. Due to its small size there is the opportunity to carry out PhD-like research projects on the course (see here for previous students? dissertations), as well having a number of bespoke taught modules with lots of opportunities to interact with the faculty and other members of the course (see here for a list of modules and here for a list of the teaching staff).

Previous members of the MPhil have gone on to study for PhDs in top research groups (e.g. at Oxford, Cambridge, Stanford, and MIT), and have gone into top industry positions (e.g. Microsoft Research, Facebook AI Research, Open AI, and AstraZeneca).

This year our programme is restructuring around four overlapping tracks: speech and language processing, computer vision & robotics, human-computer interaction, and machine learning. You apply to one of these tracks and this choice shapes your module options and the research project that you will take on. We are especially interested in candidates who are interested in speech and language processing, computer vision & robotics, and human-computer interaction as we have significant capacity to expand in these areas this year.

Details about the application process can be found on our website. The application deadline is 2nd December 2021.

Top

Le projet European Language Equality vise à établir un agenda stratégique concernant la
recherche et l’innovation pour atteindre l’égalité des langues à l’ère numérique en
Europe en 2030. Dans le cadre de ce projet, les partenaires ont produit des rapports
documentant l'état des technologies et ressources pour chaque langue officielle, ainsi
que pour certaines langues non-officielles (D1.4-D1.36). Des états de l'art couvrant
quatre grands domaines ont également été produits (D2.12-D2.16):
- la traduction automatique:
- les technologies vocales
- les technologies pour l'analyse et la compréhension des langues
- les ressources et les bases de connaissance.

Tous ces rapports sont accessibles depuis le site du projet:
https://european-language-equality.eu/deliverables/

Le rapport consacré à l'état des technologies pour la langue française et pour la langue
des signes française a fait l'objet d'une traduction en français. Il est disponible ici:
https://hal.archives-ouvertes.fr/hal-03637784

Top

Press Release – immediate
Paris, France and Tel Aviv, Israel, October 12, 2023

Distribution Agreement between ELDA and Lexicala for Multilingual Lexical Data Dissemination

ELDA and Lexicala by K Dictionaries are delighted to announce their new cooperation on distributing Language Resources for 50 languages.

ELDA is now making available Lexicala’s high-quality lexical data designed to enhance language learning, and support Machine Translation and diverse Natural Language Processing and Artificial Intelligence applications.

The Lexicala resources consist of different groups of datasets. Full descriptions can be found in the ELRA Catalogue of Language Resources under the following links:

1. GLOBAL Multilingual Lexical Data: a network of lexicographic cores for major world languages, comprising monolingual cores, bilingual pairs, and multilingual combinations for 25 languages.
2. MULTIGLOSS Multilingual Glossaries: a series of innovative word-to-sense glossaries for over 30 languages into 45 more languages.
3. Morphological lexicons: extensive morphological lists linking inflected forms to main lemmas for 15 languages.
4. Parallel Corpora & Domains: parallel corpora for nearly 400 language pairs and numerous multilingual combinations, featuring general language and vertical domain vocabularies.
5. Biographical & Geographical Names:

• English BIO Biographical Names: 4,200 dictionary entries regarding prominent persons worldwide.
• English GEO Geographical Names: 7,200 dictionary entries regarding major locations worldwide.
• GEOLINGUAL Tables: multilingual tables of over 200 countries and geographical names – including their adjectives, persons, and main languages – in 16 languages.
• Audio Pronunciation & Phonetic Transcription: human voice recordings of single-word lemmas and multiword expressions, as well as IPA and alternative scripts for 21 languages.

 For more information, please write to contact@elda.org.

 About Lexicala

Lexicala by K Dictionaries offers multi-layer lexical data for and across 50 languages, relying on 30-year experience in pedagogical and multilingual lexicography worldwide. Lexicala converges manual content creation and curation with automated data processes and helps to enhance machine translation and other natural language processing applications, as well as language learning and model training.

To find out more about Lexicala, please visit: https://lexicala.com/

About ELDA

The Evaluation and Language resources Distribution Agency (ELDA) identifies, collects, markets, and distributes language resources, along with the dissemination of general information in the field of Human Language Technologies (HLT). ELDA has considerable knowledge and skills in HLT applications. ELDA is part of major French, European and international projects in the field of HLT.

To find out more about ELDA, please visit: http://www.elda.org/

Top

Dear colleagues,

Following the success of the 2nd International Conference on Tone and Intonation (TAI 2023) <www.tai2023.org> held in Singapore 18-20 November 2023, the TAI Standing Committee is now inviting bids to host the 3rd TAI conference in 2025.

As the merger of the TAL (Tonal Aspects of Languages) and TIE (Tone and Intonation in Europe) conference series (2004–2018), the TAI series inherits features from its predecessors. The conference accepts 2-page abstracts, avoids parallel sessions, and typically hosts 3 to 5 invited keynote speakers. Following the conference, participants may choose to submit their optional 5-page full papers for publication in the ISCA online archives. With a focus on linguistic rather than technological aspects of prosody, TAI is affiliated with both the International Speech Communication Association (ISCA) and the International Phonetic Association (IPA), thereby receiving grants/awards from both organizations. TAI may also feature workshops at the same venue outside the conference program, before, after or during the conference. Given that all previous TIE conferences were held in Europe, the aim for the new TAI series is to rotate between Europe and other locations. Consequently, it is preferable for TAI 2025 to take place in Europe. While the initial two TAIs were held in November/December, the scheduling of future TAI conferences remains flexible, with spring or summer generally preferred, contingent on the preferences of local organizers.

The responsibility of the TAI Standing Committee is to promote the continuation of the series, select the venue for the next conference, and evaluate conferences and any proposals for format changes. The committee is composed of former TAL/TIE and TAI organizers. Its current members are Amalia Arvaniti, Yiya Chen, Christian DiCanio, Minghui Dong, Wentao Gu, Carlos Gussenhoven, Yanfeng Lu, Hansjörg Mixdorff, and Oliver Niebuhr. All bids will be discussed and voted on by the committee.

Bids should minimally provide the following information:

- Host institution
- General chair (and any co-organizers)
- Proposed conference dates
- Proposed conference theme
- Venue or general location of the venue
- Transportation and accommodation
- Estimated full and student registration fees
- Financial plan (including existing or potential funding sources)

Any additional information that may support the bid is welcome.

Bids should be submitted to the current chair of the TAI Standing Committee, Prof. Wentao Gu (wtgu@gavo.t.u-tokyo.ac.jp) by 30 April 2024. He is also available for additional information. Please feel free to distribute this announcement to any researchers in the speech prosody community who may be interested. Thank you very much!!

Top

• https://twitter.com/ManyLanguagesC 
• https://bsky.app/profile/manylanguagesc.bsky.social

Top

1. The Speech Prosody Conference Program is now available, at https://www.universiteitleiden.nl/sp2024/program .

2. The online lecture series resumes next month, with a talk on speech synthesis needs by Zofia Malisz; details below, also at https://sprosig.org/index.html .  After that the tentative schedule is

• Gabriel Skantze, KTH, May 15.
• Simon Roessig, York, September.
• Sam Tilsen, Cornell, October.
• Sasha Calhoun, Victoria University of Wellington, November.
• Robert Xu, Stanford, December.

The speech synthesis phoneticians need is both realistic and controllable: A survey and a roadmap towards modern synthesis tools for phonetics.
Zofia Malisz, KTH Royal Institute of Technology.
April 17th, 2 pm (Brasilia time). viewing link

ABSTRACT
In the last decade, data and machine learning-driven methods to speech synthesis have greatly improved its quality. So much so, that the realism achievable by current neural synthesisers can rival natural speech. However, modern neural synthesis methods have not yet transferred as tools for experimentation in the speech and language sciences. This is because modern systems still lack the ability to manipulate low-level acoustic characteristics of the signal such as e.g.: formant frequencies.
In this talk, I survey recent advances in speech synthesis and discuss their potential as experimental tools for phonetic research. I argue that speech scientists and speech engineers would benefit from working more with each other again: in particular, in the pursuit of prosodic and acoustic parameter control in neural speech synthesis. I showcase several approaches to fine synthesis control that I have implemented with colleagues: the WavebenderGAN and a system that mimicks the source-filter model of speech production. These systems allow to manipulate formant frequencies and other acoustic parameters with the same (or better) accuracy as e.g.: Praat but with a far superior signal quality.
Finally, I discuss ways to improve synthesis evaluation paradigms, so that not only industry but also speech science experimentation benchmarks are met. My hope is to inspire more students and researchers to take up these research challenges and explore the potential of working at the intersection of the speech technology and speech science.

Outline: 1. I discuss briefly the history of advancements in speech synthesis starting in the formant synthesis era and explain where the improvements came from. 2. I show experiments that I have done that prove modern synthesis is processed not differently than natural speech by humans in a lexical decision task as evidence that the realism (“naturalness”) goal has been largely achieved. 3. I explain how realism came at the expense of controllability. I show how controllability is an indispensable feature for speech synthesis to be adopted in phonetic experimentation. I survey the current state of research on controllability in speech engineering - concentrating on prosodic and formant control. 4. I propose how we can fix this by explaining the work I have done with colleagues on several systems that feature both realism and control. 5. I sketch a roadmap to improve synthesis tools for phonetics - by placing focus on benchmarking systems according to scientific criteria.

• TBD. Gabriel Skantze, KTH, May 15.
• TBD. Simon Roessig, York, September.
• TBD. Sam Tilsen. October.
• TBD. Sasha Calhoun, November.
• TBD, Robert Xu, December.

Nigel Ward, SProSIG Chair, Professor of Computer Science, University of Texas at El Paso

nigel@utep.edu    https://www.cs.utep.edu/nigel/   

Top