April 2025 - Corpora

Multi-Domain AI Text Detection Shared Task: M-DAIGT 2025 - Call for Participation
by Saad Ezzini 06 Apr '25

06 Apr '25

Dear Colleagues, We are pleased to announce the Multi-Domain Detection of AI-Generated Text (M-DAIGT) shared task, hosted at RANLP 2025. This task brings together researchers to explore methods for detecting AI-generated text across multiple domains, with a focus on news articles and academic writing. We invite participation in two subtasks: 1. News Article Detection (NAD): Classify news articles and snippets as human-written or AI-generated. 2. Academic Writing Detection (AWD): Identify AI-generated content within student coursework and academic research across various disciplines. * Participants will receive balanced datasets containing human-written and AI-generated texts from multiple language models. Evaluation will be conducted on the CodaLab platform. Evaluation Metrics: * Primary: Accuracy, Precision, Recall, F1-score * Secondary: Robustness across text lengths, domains, and generation sources Important Dates: * Training Data Release: March 31, 2025 * Evaluation Data Release: April 30, 2025 * Evaluation Period: May 215, 2025 * Paper Submission Deadline: May 25, 2025 * Workshop Dates: September 1112, 2025 More Information and Registration: * Website: https://ezzini.github.io/M-DAIGT/ * GitHub Repository: https://github.com/ezzini/M-DAIGT * Registration: Click here to register for solo or team participation<https://docs.google.com/forms/d/e/1FAIpQLSextZDY7qjGRJSLCBNISPcBNQZwusRWKvy…> * Join us on Slack: Slack Workspace<https://mdaigtsharedt-xye5995.slack.com/?redir=%2Fssb%2Fredirect> We look forward to your participation and encourage you to share this with colleagues who may be interested. For any queries, feel free to reach out to the organizers. Yours sincerely, The M-DAIGT Organizers ********************************************************************** DISCLAIMER: The information in this email and its attachments (if any) is intended for the addressee only and may contain confidential or privileged information. If you are not the intended recipient, please delete the email and its attachments from your system and notify the sender immediately. You should not retain, disclose, copy, or use this email or any of its contents for any purpose, nor disclose its contents to any other person. KFUPM is not responsible for changes made to this message after it was sent. Statements and opinions expressed in this e-mail are those of the sender, and do not necessarily reflect those of KFUPM. KFUPM is not liable for any effect or virus damage caused by this message. إن المعلومات الواردة في هذا البريد الإلكتروني ومرفقاته إن وجدت، قد تكون خاصة أو سرية؛ فإذا لم تكن المقصود بهذه الرسالة؛ فيُرجى منك حذفها ومرفقاتها من نظامك وإخطار المرسل بخطأ وصولها إليك فورا. كما لا يجوز نسخ أي جزء منها أو مرفقاتها ، أو الإفصاح عن محتوياتها لأي شخص أو استعمالها لأي غرض آخر. إن جامعة الملك فهد للبترول والمعادن لا تتحمل مسؤولية التغييرات التي يتم إجراؤها على هذه الرسالة بعد إرسالها. وإن البيانات أو الآراء المعبر عنها في هذا البريد، هي بيانات تخص مُرسلها، ولا تعكس بالضرورة رأي وبيانات الجامعة. كما لا تتحمل الجامعة مسؤولية أي تأثير ينتج عن هذه الرسالة أوعن أي فيروس قد تحمله.

1 0

RANLP 2025 : Join the Shared Task: Sentiment Analysis on Arabic Dialects in Hospitality
by chafik.Salmane＠um6p.ma 06 Apr '25

06 Apr '25

Sentiment Across Multi-Dialectal Arabic: A Benchmark for Sentiment Analysis in the Hospitality Domain We invite researchers, practitioners, and NLP enthusiasts to participate in the Sentiment Across Multi-Dialectal Arabic shared task, a challenge aimed at advancing sentiment analysis for Arabic dialects in the hospitality sector. About the Task Arabic is one of the world’s most spoken languages, characterised by rich dialectal variation across different regions. These dialects significantly differ in syntax, vocabulary, and sentiment expression, making sentiment analysis a challenging NLP task. This task focuses on multi-dialectal sentiment detection in hotel reviews, where participants will classify sentiment as positive, neutral, or negative across multiple Arabic dialects, including Saudi, Moroccan, and Egyptian Arabic. This shared task provides a high-quality multi-dialect parallel dataset, enabling participants to explore: 1. Dialect-Specific Sentiment Detection – Understanding how sentiment varies across dialects. 2. Cross-Linguistic Sentiment Analysis – Investigating sentiment preservation across dialects. 3. Benchmarking on Multi-Dialect Data – Evaluating models on a standardised Arabic dialect dataset. Dataset Overview - Hotel reviews across multiple Arabic dialects. - Balanced sentiment distribution (positive, neutral, negative). - Multi-Dialect Parallel Dataset – Each review is available in multiple dialects, allowing for cross-linguistic comparison. Evaluation Metrics - Primary Metric: F1-Score. - Additional Analysis: Comparison of sentiment accuracy across dialects. Baseline System - Pre-trained BERT-based model (AraBERT) fine-tuned on MSA and Arabic dialect data. - Participants are encouraged to improve upon the baseline model with their own techniques and use LLMs. Why Participate? - Contribute to Arabic NLP Research – Help advance sentiment analysis for Arabic dialects. - Gain Access to a High-Quality Dataset – A unique multi-dialect benchmark for future research. - Collaborate with the NLP Community – Engage with leading researchers and practitioners. - Showcase Your Work – High-performing models may be featured in a post-task publication. Timeline - Training data ready – April 15, 2024 - Test Evaluation starts – April 27, 2025 - Test Evaluation end – May 10, 2025 - Paper submission due – May 16, 2025 - Notification to authors – May 31, 2025 - Shared task presentation co-located with RANLP 2025 – September 11 and September 12, 2025 How to Participate 1. Register for the task via https://ahasis-42267.web.app/ 2. Download the dataset and baseline system. 3. Develop and test your sentiment analysis model. 4. Submit your results for evaluation. Organising Team - Maram Alharbi, Lancaster University, UK - Salmane Chafik, Mohammed VI Polytechnic University, Morocco - Professor Ruslan Mitkov, Lancaster University, UK - Dr. Saad Ezzini, King Fahd University of Petroleum and Minerals, Saudi Arabia - Dr. Tharindo Ranasinghe, Lancaster University, UK - Dr. Hansi Hettiarachchi, Lancaster University, UK For inquiries, please contact us at ahasis.task(a)gmail.com

1 0

PhD positions at the University of Amsterdam
by Ekaterina Shutova 05 Apr '25

05 Apr '25

PhD positions at the Institute for Logic, Language and Computation (ILLC) at the University of Amsterdam, Netherlands Salary: EUR 2.901 - EUR 3.707 gross per month Closing date: 21 April 2025 We have two open PhD positions in natural language processing (NLP), starting in September 2025 or as soon as possible thereafter. The focus of the project is on the development of methodologies for multilingual NLP and alignment of large language models. We welcome applications from candidates with an NLP / AI background and an interest in language and society. For further information and to apply: https://werkenbij.uva.nl/en/vacancies/two-phd-positions-in-natural-language… For any questions, please send an email to e.shutova(a)uva.nl

1 0

Mind the AI-GAP 2025 at HHAI 2025 – Submission Deadline extended to April 11, 2025
by bsavoldi＠fbk.eu 04 Apr '25

04 Apr '25

First “Mind the AI-GAP 2025: Co-Designing Socio-Technical Systems” International Workshop at HHAI 2025 9 June 2025, Pisa, Italy https://aigap2025.isti.cnr.it/ **Important Dates** (Time zone: Anywhere on Earth) (NEW) Submission deadline: 11 April, 2025 (NEW) Notification of acceptance: 5 May, 2025 Camera Ready due: 12 May, 2025 **Aim and scope** The Mind the AI-GAP 2025 workshop aims to critically address unwanted bias and discrimination in AI technologies by proactively integrating fairness and inclusivity within the design process, fostering social and structural change. The workshop explores how Participatory AI can shape solutions that better reflect community values, needs, and preferences and aims to bring together diverse stakeholders, including researchers, practitioners, NGOs, civil society, and designers. Through a combination of talks, roundtables, and hands-on activities, participants will collectively discuss participatory approaches and develop actionable outputs, such as guidelines or a white paper, to advance Participatory AI as a tool for equitable, transparent, and impactful systems. **Topics** We welcome technical and non-technical submissions with experimental, theoretical, or methodological contributions. We explicitly encourage interdisciplinary submissions focusing on participatory approaches to AI development. Topics of interest include, but are not limited to: Methods and frameworks for participatory AI design Case studies of co-design processes in AI development Approaches to stakeholder engagement and community value integration Analyses of power dynamics in participatory AI design Strategies for balancing individual and collective needs in AI design Methods and evaluation frameworks for participatory AI processes Tools and techniques for enhancing AI transparency for diverse stakeholders Experiences and lessons learned from co-design and stakeholder engagement Ethical considerations in participatory AI development Citizen science and democratizing AI design and deployment Real-world impacts and challenges of participatory AI design in practice The workshop is also open to other non-listed topics aligned with the scope of the venue. **Submission** We welcome the following types of submissions: - Full original research paper that presents original, impactful work (from 5 up to 9 pages); - Blue sky papers present visionary ideas to stimulate the research community (from 5 up to 9 pages); Both types of papers will be published in the conference proceedings. - Extended abstracts describing ongoing research, personal experiences with the topic, proof of concept, etc.. Authors can opt for having their paper included in the proceedings (5 pages required) or for non-archival presentations (from 2 up to 5 pages); - Research communication of already published papers that serve to promote the dissemination of contributions aligned with the scope of the workshop (up to 2 pages). They will not be published in the conference proceedings. All paper lengths exclude references, which are unlimited. All submissions should adhere to the CEUR-WS guidelines and style templates (PDF, LaTeX, Word available at https://ceur-ws.org/HOWTOSUBMIT.html) with single column format. Submissions are to be uploaded on Easychair at https://easychair.org/conferences/?conf=aigap2025. Accepted submissions shall be submitted to CEUR-WS.org for online publication in a dedicated free, open-access volume in CEUR Workshop Proceedings. Since CEUR partners with Scopus, these proposals will also be indexed in it. Contributions will be presented either as oral presentations (lightning talks) or posters. All presentations are expected to be in person, except in exceptional cases (e.g., a speaker encounters a last-minute issue and cannot attend the conference). **Workshop organizers** Costanza Alfieri, Università dell’Aquila Eleonora Cappuccio, Consiglio Nazionale delle Ricerche Donatella Donati, Università dell’Aquila Miriam Felici, Independent Researcher Marta Marchiori Manerba, Università di Pisa Benedetta Muscato, Scuola Normale Superiore Clara Punzi, Scuola Normale Superiore Beatrice Savoldi, Fondazione Bruno Kessler For more information: Website: https://aigap2025.isti.cnr.it/ Contact: mind-the-ai-gap(a)googlegroups.com

1 0

Call for Participation in SHARED TASK: Analysis of Persuasion Techniques — in parliamentary debates and propaganda-oriented social media
by Roman Yangarber 04 Apr '25

04 Apr '25

*Call for Participation in Shared Task* ** Analysis of Persuasion Techniquesin Parliamentary Debates and Disinformation- and Propaganda-oriented Social Media * Co-located with Slavic NLP 2025 <http://bsnlp.cs.helsinki.fi/>Workshop, at ACL in Vienna, Austria bsnlp.cs.helsinki.fi/shared-task.html <http://bsnlp.cs.helsinki.fi/shared-task.html> * * TASK DESCRIPTION: * * The task focuses on detection and classification of Persuasion Techniques using data from 5 Slavic languages — Bulgarian, Polish, Croatian, Slovene and Russian — in two types of texts: (a) parliamentary debates on hotly-contested topics, and (b) social media posts, related to the spread of disinformation and propaganda. The task has two subtasks: 1. Subtask 1: Detection — Given a text and a list of fragment offsets, determine for each fragment whether it contains one or more persuasion techniques, from a given taxonomy of persuasion techniques, 2. Subtask 2: Classification —Given a text and a list of fragment offsets, determine for each fragment which persuasion techniques are employed therein. We use a rich taxonomy with 25 persuasion techniques: Name-calling or labelling, Guilt by association, Casting doubt, Appeal to hypocrisy, Questioning the reputation, Flag waiving, Appeal to authority, Appeal to popularity, Appeal to fear and prejudice, Appeal to values, Strawman, Whataboutism, Red herring, Appeal to pity, Causal oversimplification, False dilemma or no choice, Consequential oversimplification, False equivalence, Slogans, Conversation killer, Appeal to time, Loaded language, Obfuscation-Intentional vagueness-confusion, Exaggeration or minimization, Repetition. Subtask 1 is a binary classification task. Subtask 2 is a multi-class multi-label classification task. The text fragments correspond to paragraphs. For information about training and test data, guidelines, and participation, please see theShared Task Home Page. <http://bsnlp.cs.helsinki.fi/shared-task.html> IMPORTANT: Participants may join both subtasks or only one. It is not mandatory to submit responses for all languages. Up to 5 system responses per language per team may be submitted. Important Dates * Registration deadline: 26 April 2025 * Release of Testdata to registered participants: *29 April*2025 * Submission of system responses: 5 May 2023 * Results announced to participants: 8 May 2025 * Submission of shared task papers (optional): 18 May 2025 * ** *Questions and contact: bsnlp(a)cs.helsinki.fi<mailto:bsnlp@cs.helsinki.fi>* -- Roman Yangarber Professor, University of Helsinki, Finland Digital Humanities INEQ: Helsinki Inequality Initiative <https://helsinki.fi/en/ineq-helsinki-inequality-initiative> — Linguistic Inequalities and Translation Technologies ------------------------------------------------------------------------ e-Learning & language learning Language Learning Lab Unioninkatu 40, Metsätalo A214 revitaAI.github.io <https://revitaai.github.io> helsinki.fi/language-learning-lab <https://www.helsinki.fi/language-learning-lab> mobile: +358 50 41 51 71 3 ------------------------------------------------------------------------ RЯ <https://www.helsinki.fi/language-learning-lab>

1 0

Shape the Future of AI-Powered Wearables – Join the Meta CRAG-MM Challenge! 🚀
by Kai Sun 04 Apr '25

04 Apr '25

Hello Colleagues, Imagine you're shopping for dinner at your favourite grocery store wearing AI glasses. You scan the shelves looking for the right ingredients. You ask, “Which of these pasta is gluten-free?”—and in seconds, your AI assistant provides an accurate answer, cross-referencing product labels and trusted sources. No guesswork. No misinformation. Just reliable, real-time knowledge at your fingertips. This is the potential of AI-powered wearable devices. But today’s Vision Large Language Models (VLLMs) still struggle with providing accuracy, context, and real-time information. That’s where you come in. The Meta CRAG-MM Challenge for KDD Cup 2025 is pushing the boundaries of multi-modal, multi-turn Retrieval-Augmented Generation (RAG) for wearable AI. Join the challenge to build AI that sees, understands, and retrieves knowledge—without hallucinating. Why This Challenge Matters Despite advances in AI, VLLMs still generate hallucinated answers—especially when handling long-tail knowledge, multi-step reasoning, or real-world images. The Meta CRAG-MM Benchmark introduces a rigorous test for multi-modal retrieval and reasoning, helping to ensure AI models can accurately process images, retrieve external information, and handle smooth, multi-turn conversations. What Makes CRAG-MM Unique? 🔍 Real-world wearable AI focus – Uses real-world images from Ray-Ban Meta glasses across 14 diverse domains 🤖 Multi-modal + Multi-turn challenge – Evaluate AI across single-turn and multi-turn question-answering 🧠 Complex question types – Tests reasoning, aggregation, and retrieval beyond simple fact lookup 💰 $33,000 in prizes – Compete for the grand prize and recognition at KDD Cup 2025. Challenge Timeline 📅 Warm up phase begins March 24, 2025 – Open to all teams 📅 Phase 1 begins Apr 5, 2025 – Open to all teams 📅 Phase 2 begins May 11, 2025 – Top-performing teams advance 🏆 Winners announced August 5, 2025 at KDD Cup 2025 in Toronto Get Started Today! 👉 Register now: <https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025> https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025 💬 Join the community: Connect with others on our <https://discourse.aicrowd.com/c/2929> Community Forum<https://discourse.aicrowd.com/c/2929> & <https://discord.gg/YWDQQa8byx> Discord<https://discord.gg/YWDQQa8byx> 📜 Read the rules: <https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025/challenge_ru…> https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025/challenge_ru… All The Best, Team AIcrowd NO PURCHASE NECESSARY TO ENTER/WIN. Open to individuals who are 18 + age of majority and meet the full eligibility requirements in the full Rules. Open 3/17/2025 23:55 UTC through 6/1/2025 23:55 UTC. Void where prohibited. Subject to full Rules at <https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025/challenge_ru…> https://www.aicrowd.com/challenges/meta-crag-mm-challenge-2025/challenge_ru…. See Rules for prize details and values. Sponsor: Meta Platforms, Inc. 1 Hacker Way, Menlo Park, California 94025 (for US entrants) or Meta Ireland Limited, 4 Grand Canal Square, Dublin 2, Ireland (for all other entrants).

1 0

Second Call for Participation: Shared task on Clickbait Detection and Spoiling in Spanish - TA1C
by Luis Chiruzzo - Inco 04 Apr '25

04 Apr '25

*** Second Call for Participation for TA1C at IberLEF 2025 *** TA1C (Te Ahorré Un Click) Clickbait Detection and Spoiling in Spanish at IberLEF 2025 https://codalab.lisn.upsaclay.fr/competitions/21819 UPDATE: The training data has been released! Clickbait is a widespread phenomenon in online news: it is a way of creating headlines and teasers aimed at capturing readers’ attention in order to increase traffic, relegating the function of informing to a secondary role. There is no clear consensus at the moment about how to define clickbait exactly, with some contradictory definitions that usually are based on the deceptive effect created by the news failing to deliver what they promise, or content based related phenomena such as sensationalism or yellow journalism. For this task we will take the following definition, based on Loewenstein's information gap theory: “Clickbait is a method for generating teasers, especially online, that deliberately omits part of the information with the goal of generating curiosity by creating an information gap, thereby attracting the readers' attention and making them click”. Although clickbait started in low-reputation web-exclusive media that focused on political propaganda or soft-news, such as The Huffington Post, Buzzfeed and Upworthy, it has gained prominence across all types of news and media. However, it is usually perceived as annoying and it can lead to misinformation. Spoiling the clickbait involves satisfying the curiosity by answering the information gap created. This way, the reader could have all of the information and can decide to read the complete article based on interest and not curiosity, just as if the headline was written in a traditional way. In this shared task we will provide a dataset of media tweets written in different varieties of Spanish and from different sources, with their corresponding associated media articles. Participants will be asked to solve the following tasks: * Clickbait Detection: Determine if the content of a tweet that links to a media article is clickbait, given the previous definition of clickbait. This is a binary classification task. * Clickbait Spoiling: Given a clickbait teaser (tweet and title) and the corresponding news article, generate or extract from the article a short text that, as concisely as possible (280 characters max), fills the information gap, satisfying the generated curiosity, or otherwise indicate that the articles has no response for it. The generated text must be in Spanish. How to participate: If you want to participate in this task, please join our Codalab competition <https://codalab.lisn.upsaclay.fr/competitions/21819>: Important Dates: * May 27th, 2025: test set and open for submissions. * June 3rd, 2025: publication of results. * June 12th, 2025: paper submission. * June 20th, 2025: notification of acceptance. * June 27th, 2025: camera-ready paper submission. * September, 2025: IberLEF 2025 Workshop.

1 0

Second CFP-Fifth Workshop on Language Technology for Equality, Diversity, and Inclusion (LT-EDI-2025)
by Bharathi Raja Asoka Chakravarthi 03 Apr '25

03 Apr '25

Fifth Workshop on Language Technology for Equality, Diversity, and Inclusion (LT-EDI-2025) To be held at the 5th Conference on Language, Data and Knowledge (LDK 2025) We’re happy to invite paper submissions for the LT-EDI 2025 workshop—an important platform for exploring how Language Technology (LT) can drive equality, diversity, and inclusion across languages, cultures, and communities. Call for Papers: Topics of interest include, but are not limited to: - Related to speech and language resource creation for EDI. - Data set development to include EDI. - Gender inclusivity in LT. - LGBTQ+ inclusivity in LT. - Racial inclusivity in LT. - Persons with disability' inclusivity in LT. - Speech and language recognition for minority groups. - Unconscious bias and how to avoid it in Natural Language Processing, Machine Learning, and other applications of LT. - Tackling rumors and fake news about gender, racial, and LGBTQ+ minorities. - Tackling discrimination against gender, racial, and LGBTQ+ minorities. - Counter-narrative applied to LGBTQ+ minorities. Important Dates Paper Submission Deadline: 15 May 2025 Notification of Acceptance: 12 June 2025 Camera-ready Paper Due: 26 June 2025 Workshop Date (Tentative): 9 September 2025 Website: https://sites.google.com/view/lt-edi-2025/home?authuser=0 🌐 Submission Link: https://openreview.net/group?id=LDK/2025/Workshop/LT-EDI with regards, Dr. Bharathi Raja Chakravarthi, Assistant Professor / Lecturer-above-the-bar Programme Director (MSc Computer Science - Artificial Intelligence) <https://www.universityofgalway.ie/courses/taught-postgraduate-courses/compu…> School of Computer Science, University of Galway, Ireland Insight SFI Research Centre for Data Analytics, Data Science Institute, University of Galway, Ireland E-mail: bharathiraja.akr(a)gmail.com , bharathi.raja(a)universityofgalway.ie <bharathiraja.asokachakravarthi(a)universityofgalway.ie> Google Scholar: https://scholar.google.com/citations?user=irCl028AAAAJ&hl=en Website: https://www.universityofgalway.ie/our-research/people/computer-science/bhar… <https://www.universityofgalway.ie/our-research/people/computer-science/bhar…>

1 0

Call for Papers - NLP4Sustain Workshop 2025
by Jakob, Charlott Frederike 03 Apr '25

03 Apr '25

Call for Papers: NLP for Sustainability (NLP4Sustain) Workshop 2025 With this workshop, we want to provide an interdisciplinary forum for discussing research, progress, and challenges in the context of NLP and sustainability. We invite submissions about NLP-based analysis of sustainability-related texts, sustainable NLP models and evaluation practices in general. Authors and other participants will engage with each other in a poster session and there will be an interdisciplinary invited talk with an ensuing discussion. We invite technical, survey, and position papers, as long (8 page) or short (4 page) papers (plus references and appendices) written in English and formatted according to the ACL stylesheet. Relevant Topics • analyses and classifications of sustainability-related texts (such as company reports, advertisements, legal texts, …) • generation of explanations, critiques, summaries, … of sustainability-related texts • multimodal models related to sustainability, such as language-vision or climate-impact models • question answering in the sustainability/climate context • sustainable (e.g. small, efficient) NLP models for other applications/domains • sentiment analysis in the sustainability/climate context • media and social media analysis with NLP methods in the sustainability/climate context Important Dates * Tue, 10.06.2025: Paper submission deadline * Fri, 25.07.2025: Acceptance notification * Fri, 15.08.2025: Camera-ready due * Wed, 10.09.2025: Workshop @KONVENS in Hildesheim, Germany (at least one author must be present) For details, visit the website: https://nlp4sustain.github.io/ If you have any questions, please contact: jakob.prange(a)uni-a.de<mailto:jakob.prange@uni-a.de> and/or c.jakob(a)tu-berlin.de<mailto:c.jakob@tu-berlin.de> —————————————————— Charlott Jakob (she/her) Academic Researcher Technische Universität Berlin Quality and Usability Lab Institute of Software Engineering and Theoretical Computer Science Faculty IV Electrical Engineering and Computer Science Technische Universität Berlin Sekr. MAR 6-7, Marchstr. 23, 10587 Berlin, Germany

1 0

Post-doc position in computational linguistics at Univ. Lorraine (France)
by Mathieu Constant 02 Apr '25

02 Apr '25

The Research unit ATILF (Computer Processing and Analysis of the French Language) offers a postdoctoral position in computational linguistics. Topic: multiword expressions in large language models Location: ATILF, Nancy, France (Univ. Lorraine and CNRS) Starting date: September 2025 Duration: 12 months (possibility to extend the duration for one more year) Supervisors: Mathieu Constant (Univ. Lorraine, France) and Patrick Watrin (UC Louvain, Belgium) Salary: depends on experience and salary grids (from 3000 to 4200 euros before tax) Application deadline: April 22, 2025 Subject. The term « multiword expression » (MWE) refers to a combination of multiple lexical items that displays irregular composition possibly on different linguistic levels (morphology, syntax, semantics, …). They include a large variety of phenomena such as idioms (run around in circles), support verb constructions (take a walk), nominal compounds (dry run), complex function units (in spite of). They have been the subject of extensive research work in the NLP community over the last 50 years. The goal of this post-doc position is to investigate to what extent large language models encode multiword expressions and their various levels of idiomaticity and fixedness. In particular, the hired post-doc will develop methods to extract linguistic features about multiword expressions in context from large language models. The methods will be experimented on French and will be used to provide aids for French L2 learners when reading MWE occurrences in authentic texts. Context. The position is part of the STAR-FLE project (STrategic Adaptations for better Reading and Text Comprehension in FFL, https://www.starfle.fr/en, 2024-2027) funded by the French National Research Agency (ANR). The project aims to propose innovative digital solutions in the area of Natural Language Processing (NLP) that may improve text comprehension for French L2 learners and assist teachers in managing multiple levels of learners. In particular, it will propose context-based aids for understanding lexical issues as well as MWEs found in authentic texts. The hired researcher will be fully integrated in the project team. Requirements. Applicants should hold a PhD thesis n natural language processing, in computational linguistics, in computer science, or in applied mathematics, . The hired post-doc researcher should have the following skills: * expertise in deep learning for NLP and notably large language models * excellent programming skills * Good linguistic skills * good knowledge of French would be a plus * team spirit Application. The applicants should submit a coverage letter, a CV including their publications, a list of references for recommandation, on the following official web site: https://emploi.cnrs.fr/Offres/CDD/UMR7118-SABMAR-022/Default.aspx?lang=EN. The applications should be sent not later than April 22, 2025. For more information, contact Mathieu Constant (Mathieu.Constant(a)univ-lorraine.fr)

1 0

2025

2024

2023

2022

Corpora April 2025