SIGUL July 2026

sigul@list.elra.info

5 participants
5 discussions

Re: join ISCA SIGUL
by Laurent 22 Jul '26

22 Jul '26

dear anna tks for your interest in SIGUL i’m no longer involved in it but people in cc are best L > On 22 Jul 2026, at 14:56, Anna Taylor <anna.taylor(a)eurecom.fr> wrote: > > Dear Laurent, > > I would like to join the ISCA SIGUL as I am a PhD student in speech deepfake detection with an interest in lesser-resourced languages. > > Thank you, > > -- > Anna TAYLOR > > EURECOM Doctorante > EDITE PhD Student Representative > Digital Security Department > Sophia Antipolis, France > Office 460-3 > www.eurecom.fr <https://www.eurecom.fr/> >

1 0

IWSLT 2027 - Call for Task Proposals
by Atul K. Ojha 22 Jul '26

22 Jul '26

Dear all, The International Conference on Spoken Language Translation (IWSLT) <https://iwslt.org/> is the premier annual conference for all aspects of Spoken Language Translation. Every year, the conference organizes and sponsors open evaluation campaigns around key challenges in simultaneous and consecutive translation, under real-time/low latency or offline conditions, and for a variety of languages in under-resourced or multilingual conditions. System descriptions and results from participants’ systems, as well as scientific papers on key algorithmic advances and best practices, are presented. IWSLT is the venue of the SIGSLT, the Special Interest Group on Spoken Language Translation of ACL, ISCA, and ELRA <https://iwslt.org/>. With a track record of 20+ years, IWSLT benchmarks and proceedings serve as references for all researchers and practitioners working in speech translation and related fields. Many challenges in speech translation have not yet been addressed; among them, we are really interested in topics related to new application scenarios (e.g. meetings, subtitling, dubbing), specific aspects (e.g. names, accents), different styles, multilingually, discourse and summarization, multimodal and multi-party speech translation, and automatic evaluation metrics for speech translation. or many other ideas that researchers have not yet focused on. Therefore, we invite proposals for shared tasks. As a task organizer you can promote a particular challenge in speech translation, either newly identified or worthy of continued research. For those proposing new tasks, for inspiration, you can find the tasks run in the previous year on the IWSLT website <https://iwslt.org/2026/>. Tasks will be selected in November based on their relevance and readiness for the evaluation campaign, to enable data to be released by the end of the year. To ensure an appropriate number of total tracks, highly related proposals may be encouraged to merge after initial review. If you want to propose a new task to encourage researchers around the world to work on particular timely challenges in SLT, please fill out the following form <https://iwslt.org/assets/pdfs/IWSLT2027-Call_for_Tasks.pdf> and send it to: *iwslt-organizers(a)googlegroups.com <iwslt-organizers(a)googlegroups.com> *by* September 30th, 2026. *Decisions about which tasks will run in 2027 will be announced by *November 1st, 2026* . For further information on this initiative, please refer to the * <https://iwslt.org/assets/pdfs/IWSLT2026-Call_for_Tasks.pdf>CFP <https://iwslt.org/assets/pdfs/IWSLT2027-Call_for_Tasks.pdf> *( https://iwslt.org/assets/pdfs/IWSLT2027-Call_for_Tasks.pdf). Best, Marcello, Alex, Antonios, Luisa, Matteo, Marco, Jan, Sebastian, Elizabeth, Atul IWSLT Organisers

1 0

CfP: 10th Workshop on the Use of Computational Methods in the Study of Endangered Languages (ComputEL-10)
by Antti Arppe 21 Jul '26

21 Jul '26

ComputEL-10: Tenth Workshop on the Use of Computational Methods in the Study of Endangered Languages First CALL FOR PAPERS Workshop dates: March 2-3, 2027 Place: Honolulu, Hawaii, USA (hybrid) Submission deadline: October 2, 2026 Submission link: TBD ComputEL-10 will be co-located with the 10th International Conference on Language Documentation and Conservation (ICLDC 10) in Honolulu, Hawaii. ComputEL-10 and ICDLC 10 do **not** have joint registration. If you wish to attend ICDLC 10, visit their website: https://ling.lll.hawaii.edu/sites/icldc/. ComputEL-10 explores the interface and intersection of computational linguistics, documentary linguistics, and community-based efforts in language revitalization and reclamation. We encourage submissions that present their work from an interdisciplinary perspective with a focus on one of three areas: (i) Academic: develop new methods or technologies for tasks or applications focused on low-resource, and in particular, endangered languages; propose models in the study or processing of under-documented languages; or evaluate tools or methods for curation and analysis of data in such languages, oriented towards language documentation, description, reclamation and revitalization; (ii) Community: develop new technology toward the goals of language reclamation, revitalization, including language teaching and acquisition; propose new models for the collection, management, and mobilization of language data in community settings, with attention to issues of data sovereignty and community protocols and involvement; describes new uses of language technology by the language's community; (iii) Collaboration: describe challenges and rewards in past or on-going collaboration and interaction among computer scientists and/or documentary linguists that involve(d) language communities; formulate concrete steps for a more fruitful interaction among computer scientists, documentary linguists, and language communities. Note: priority will be given to descriptions of actual collaborations (successful or unsuccessful) over proposals or position papers. (For an example, see Cox et al. 2025: https://aclanthology.org/2025.computel-main.12/) IMPORTANT DATES 2 October, 2026 Deadline for submission of papers 13 November 2026 Notification of Acceptance 2-3 March 2027 Workshop SUBMISSIONS We offer two types of submissions: abstracts and papers. Please carefully read the requirements below. Submissions that do not fulfill all relevant requirements will be desk-rejected. The type of submission does not influence the likelihood of acceptance. GENERAL REQUIREMENTS FOR ALL SUBMISSIONS All submissions must be anonymous. Both submission types have the same deadline. Authors may submit no more than two submissions where they are co-author. In no case may an author submit more than one individually- or first-authored submission. Authors must submit and present in English. Proposals will undergo blind review, so please do not list your name, affiliation, or contact information within the file. (Author names appearing in reference lists or citations are fine and do not need to be anonymized.) ABSTRACTS Abstracts clearly describe the content that author intends to present to the ComputEL-10 audience. Abstracts are max. 800 words not including title or references. Abstracts will NOT be published (in the Proceedings of the workshop in ACL Anthology). Abstracts are intended for academics not accustomed to publishing in proceedings or non-academic community members wishing to share with and receive feedback from the other communities. Upload your abstract as a PDF or MS Word file. PAPERS Paper submission can be of two lengths: short (up to 4 pages) or long (up to 8 pages) paper. The length of submission does not influence the likelihood of acceptance. Both paper types must include a section on ethical consideration and a section on limitations; these sections are not considered part of the page limit. Papers must follow the style and formatting guidelines provided in by ACL Style Files (download template files for LaTeX: https://github.com/acl-org/acl-style-files). Submissions that violate anonymity, exceed the length requirements, or are missing an ethics or a limitations section will be desk rejected. ACL has discontinued making available a submission template for Microsoft Word. Therefore, we strongly recommend using the above LaTeX template with Overleaf, which enables WYSIWYG editing. Nevertheless, we do allow the use of the older ACL Microsoft Word template: https://2023.aclweb.org/calls/style_and_formatting/. While we intend to help accepted authors convert their paper into a format acceptable for ACL Anthology, we may not be able to resolve all potential discrepancies. Therefore, using the MS Word template may run the risk of not being accepted for publication. If you wish assistance with LaTeX, please email soon. A. Short Paper: Short paper submissions must describe original and unpublished work. They are max. 4 pages excluding references and sections on ethical consideration and limitations. A short paper is not a shortened long paper. Instead, short papers have a small, focused contribution or describe work in progress (“working paper”). Recommended topics for short papers include pilot studies, negative results, opinion pieces, interesting application nuggets, or descriptions of initial collaborative teamwork. B. Long Paper: Long papers must describe substantial, original, completed and unpublished work. Long papers are max. 8 pages excluding references and appendices and sections on ethical consideration and limitations. Wherever appropriate, concrete qualitative and quantitative evaluation and analysis should be included. PROCEEDINGS The Organizing Committee will select from the papers (excluding all abstracts) that have been accepted for presentation at the workshop for online publication via the open-access ACL Anthology. Not all accepted papers for presentation are guaranteed inclusion in the publication. Long and short papers that are accepted for publication will be allotted one additional page (altogether 5 and 9 pages) excluding references and ethics and limitations sections. Papers accepted for inclusion in the Anthology must be revised versions of the work that underwent ComputEL-10 review. Revisions and improvements should address reviewer comments or add relevant details and clarifications, but should not include entirely new, unreviewed content. ADDITIONAL AND CONTACT INFORMATION Please see the ComputEL-10 website for further information: https://computel-workshop.org/ComputEL-10/ Organizing Committee Email: computel.workshop(a)gmail.com -- ====================================================================== Antti Arppe - Ph.D (General Linguistics), M.Sc. (Engineering) Professor of Quantitative Linguistics Director, Alberta Language Technology Lab (ALTLab) Project Director, 21st Century Tools for Indigenous Languages (21C) Department of Linguistics, University of Alberta Algonquian Studies Association - Secretary-Treasurer E-mail: arppe(a)ualberta.ca - antti.arppe(a)iki.fi WWW: www.ualberta.ca/~arppe - altlab.ualberta.ca - 21c.tools Mānahtu ina rēdûti ihza ummânūti ihannaq - dulum ugulak úmun ingul ----------------------------------------------------------------------

1 0

ALPS 2027 Winter School: 1st Call for Participation
by Laurent Besacier 10 Jul '26

10 Jul '26

ALPS 2027: The seventh Advanced Language Processing School Dates: Sunday 28th March to Friday 2d April, 2027 Location: Aussois, French Alps Website: https://lig-alps.imag.fr/ <https://lig-alps.imag.fr/> About ALPS ALPS is co-organized by LIG (Univ. Grenoble Alpes), Naver Labs Europe, and Poolside.ai, and consists in a week-long series of lectures by world-class NLP researchers, sessions where participants present their work, social sessions, lab sessions, and outdoor activities in the mountains. The school will take place in the Alps (the Vanoise massif) at Aussois (1500m) and close to France’s first national park. What ALPS represents: Advanced lectures by first class researchers. An atmosphere that fosters connections and interactions. A poster session for attendees to present their work, gather feedback and brainstorm future work ideas. Target audience The intended audience of ALPS 2027 are graduate students (PhD students or advanced master students) in natural language processing or related fields. We also welcome other NLP practitioners working in industry or academia. Application: will open soon but instructions are here https://lig-alps.imag.fr/index.php/application/ <https://lig-alps.imag.fr/index.php/application/> You will need a resumé (maximum length: 2 pages) and a cover letter. For the cover letter, it should be in English and explain your motivation in attending this advanced research school. We are unfortunately not capable of accepting all candidates, and this letter is a key component that helps us to ensure that the participant’s profiles are balanced and diverse.

1 0

sharing information about BabyHuBERT
by Alex CRISTIA 06 Jul '26

06 Jul '26

Dear all, Over the past four years, researchers worldwide have entrusted the ExELang team <http://exelang.fr/> led by Alex Cristia with their raw audio recordings (either through bilateral Data Transfer Agreements/Memoranda of Understanding, or through scientific archives), which allowed the development of several computational models. Today, we are writing to you about one called BabyHuBERT. BabyHuBERT is a speech representation model, a sort of “jack of all trades” that can then be specialized to solve a specific task, such as voice type classification. The release of this model raised some important ethical questions. To get a non-technical introduction to BabyHuBERT and learn about our approach to its ethics, check out the slides <https://osf.io/n75rs/files/3hqzv> and video recording <https://osf.io/n75rs/files/jaume> (also available on YouTube <https://www.youtube.com/watch?v=inWK5YVET9Q&t=2s>) of a talk Alex gave recently for some researchers who had shared their data with us. The technical description of the model has been reported on "BabyHuBERT: Multilingual Self-Supervised Learning for Segmenting Speakers in Child-Centered Long-Form Recordings" (https://arxiv.org/abs/2509.15001), recently accepted at Interspeech. For those interested in potentially using the model, we invite you to visit this OSF Link <https://osf.io/n75rs/overview>, which includes: - A PDF with the documents surrounding the decisions taken regarding the release <https://osf.io/n75rs/files/964df> - The BabyHuBERT License v1.0 <https://osf.io/b79v4>- An introduction to BabyHuBERT <https://osf.io/n75rs/files/anx9v>, a FAQ <https://osf.io/n75rs/files/u2nrf>, and a process flow document <https://osf.io/n75rs/files/a7t2y> Please feel free to reach out to Marvin Lavechin <marvinlavechin(a)gmail.com> if you have any questions! Best, The ExELang team --------------------------------------------------------------- Alex (Alejandrina) Cristia --------------------------------------------------------------- If you donate, ask me about effective charities <https://effectivealtruism.us8.list-manage.com/track/click?u=52b028e7f799cca…>. / Si vous faites des dons, nous pouvons échanger sur le don efficace <https://mieuxdonner.fr/>.

1 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

SIGUL July 2026