Birds of Feather & Affinity Group Events Schedule
Birds of Feather Sessions will take place during the Main conference Days on Wednesday, March 25 - Friday, March 27, 2026.
Please also check the online conference schedule on Underline/Whova for the latest updates.
Wednesday, March 25, 2026
Time: 11:30 – 13:00
Arabic NLP
Organized by: Go Inoue, Salam Khalifa
Location: Salle Lamrissa Level 2 (stairs)
Description: This session brings together researchers and practitioners working on Arabic NLP. It provides an opportunity to connect, share ideas, discuss recent advances and challenges, and explore career paths and collaborative initiatives. Everyone currently working on, interested in, or simply wanting to learn more about Arabic NLP is warmly welcome to join.
Emotion Analysis and Generation
Organized by: Lynn Greschner and Roman Klinger
Location: Salle Les Oudayas Level 2 (stairs)
Description: Emotion analysis in NLP is the field of studying how affect and related concepts are expressed in text, and how texts are emotionally perceived. Such research includes aspects of stimulus detection and emotion experiencer detection, emotion classification and regression, studying communication and the lack of information from a sender to a receiver (including perspectivism), as well as psychological theories of affect. In this meeting we bring together researchers from NLP and adjacent areas to facilitate entering the field of emotion analysis and computational affective sciences.
Time: 13:30 – 14:15
SomosNLP
Organized by: Selene Baez
Location: Salle Lamrissa Level 2 (stairs)
Description: SomosNLP is a community dedicated to creating and sharing resources that support the development of NLP in Iberoamerican languages. Iberoamerica includes countries in the Americas where Spanish or Portuguese is spoken, as well as Spain and Portugal. Our goal is to connect with people from the region, support each other, and explore solutions to common challenges. We especially welcome speakers of less-represented languages to be part of this vibrant, multilingual dialogue.
We will begin the session with a short introduction to the community, its mission, vision, and ongoing open initiatives. Then, participants will be invited to introduce themselves, highlighting their work and challenges faced. This will lead to an open discussion of possible solutions and collaborations. We will wrap up by summarizing key takeaways, setting next steps, and taking a group picture to mark the occasion.
The session will take place during the lunch break. We will take care of the reservation and bring stickers, but please note that food is not sponsored.
Join us! More info at http://somosnlp.org/conferencias/eacl-2026
Time: 14:30 – 16:00
EquiCL
Organized by: Zeerak Talat
Location: Salle Les Oudayas Level 2 (stairs)
Description: Given geopolitical changes, diversity initiatives are at a crossroads and the path forward is unclear. In this session, we will hold an introduction to EquiCL and discuss the challenges that diversity initiatives are currently facing, pathways forward, and what will be required.
Time: 16:30 – 18:00
Learning and Reasoning for Structured Data
Organized by: Vivek Gupta
Location: Salle Les Oudayas
This Birds of a Feather session brings together researchers and practitioners interested in how learning and reasoning systems operate over structured and semi-structured data, including tables, charts, graphs, maps, forms, and other structured representations. The session will focus on challenges and opportunities at the intersection of machine learning (particularly LLMs and multimodal models) and structured reasoning, such as schema awareness, symbolic execution (e.g., SQL, constraints, graphs), hybrid neural–symbolic methods, and faithful reasoning and evaluation over complex structures.
We welcome discussion on representation learning for structured data, benchmark and evaluation design, robustness and faithfulness of reasoning, and real-world applications where structure is central (e.g., science, healthcare, finance, policy, and education). The goal is to surface shared open problems, compare emerging approaches, and foster cross-community collaboration spanning NLP, ML, data management, and HCI.
Time: 16:30 – 18:00
Multilingual and Multimodal Cultural Inclusivity in LLMs
Organized by: Firoj Alam
Location: Salle Lamrissa Level 2 (stairs)
Description: The Birds of a Feather (BoF) session on “Multilingual and Multimodal Cultural Inclusivity in LLMs” will bring together researchers and practitioners to discuss how to build and evaluate LLMs that better reflect diverse languages, dialects, cultures, and modalities (text, speech, and vision). We will share challenges and emerging solutions around data gaps, cultural and linguistic bias, code-switching and dialect variation, and culturally grounded evaluation. The session will be interactive, aiming to identify practical next steps and collaboration opportunities, such as shared datasets, benchmarks, and community guidelines, to make LLMs more inclusive and equitable.
Thursday, March 26, 2026
Time: 09:00 – 10:30
Muslims in ML
Organized by: Ehsaneddin Asgari
Location: Salle Les Oudayas Level 2 (stairs)
Description: The Muslims in Machine Learning (MusIML) Affinity Group Meeting provides a welcoming, identity-affirming space for individuals who identify as Muslim or are culturally affiliated with Muslim communities within the NLP and broader AI research ecosystem. This session offers a platform for community-building, mutual support, and meaningful dialogue around shared experiences in academia and industry. It also fosters discussions on the intersections of Muslim identity, ethical AI, and linguistic diversity, particularly relevant to the global, multilingual, and multicultural focus of NLP. Open to members and allies, the session will include networking, mentorship opportunities, and space to explore shared challenges and aspirations within the field.
MusIML is an independent, grassroots initiative committed to advancing equity, representation, and collaboration for Muslims in AI/ML. The organization has hosted three successful workshops at NeurIPS (2020, 2023, 2024, 2025) and ICML 2025, and it regularly organizes social and mentoring events at major conferences, including ACL 2025, ICLR 2025, and AISTATS 2025. MusIML has previously partnered with industry organizations such as Google and Amazon for past events and aims to further expand industry partnerships in the future to support and grow these community-building efforts. MusIML maintains a vibrant Slack community of close to 400 members. More details can be found at: www.musiml.org.
Time: 9:00 - 10:30
Ask us Anything - Conference 101
Organized by: EACL 2026 D&I Committee
Location: Salle Lamrissa Level 2 (stairs)
Our mentorship program is officially in motion, and we want to ensure you feel fully prepared for the week ahead. Join the D&I committee for a low-pressure AMA (ask me anything) designed to help you navigate the conference with confidence. Whether you’re wondering how to structure your first 1-on-1 mentor meeting, how to approach a senior researcher in the hallway, or how to handle specific application hurdles, we’re here to help. Why join us?: Get advice from a panel that offers a broad range of perspectives. Our members are currently active at every stage of the academic career path, including PhD students, postdocs, early-career researchers, and faculty. This allows us to provide immediate, practical advice on navigating the current landscape while also offering long-term insights into professional development.
Time: 11:00 – 12:30
Biomedical Retrieval and RAG: In the Lab and Beyond
Organized by: Vera Pavlova
Location: Salle Les Oudayas Level 2 (stairs)
Description: Biomedical retrieval and Retrieval-Augmented Generation (RAG) have become critical components of modern biomedical NLP systems. As their usage increases, limitations become more visible.
This BoF brings together researchers and practitioners working on biomedical retrieval and RAG, spanning both lab-based model development and real-world deployment. We will discuss challenges across the pipeline, including model development, domain adaptation, and evaluation gaps, as well as deployment constraints such as model size, robustness, and system behavior in practice.
The goal is to discuss shared pain points and surface insights on what improvements can help bring better biomedical retrieval models and RAG systems into real-world use, fostering dialogue, mutual feedback, and closer alignment between research and deployment.
Scientific Publishing: Stop Pretending This Works
Organized by: Mark Cieliebak
Location: Salle Lamrissa Level 2 (stairs)
Description: Our research ecosystem is broken. Conferences are bloated and exclusionary. Peer review is overloaded and unsatisfying. Publishing pipelines move at the speed of bureaucracy, not discovery. And yet, as a community, we keep acting as if the system is fine — or at least “good enough.”
It isn’t.
“Scientific Publishing: Stop Pretending This Works” is an open and informal forum for researchers who know the current model is unsustainable and are ready to talk about what comes next. We’re not here to present a polished solution. We’re here to question assumptions, surface frustrations, and explore radically different possibilities for how science could be shared, evaluated, and advanced.
This is a space for uncomfortable truths, bold ideas, and collective imagination. If you’ve ever thought “there has to be a better way,” this is your invitation to help shape it.
Time: 13:00 – 13:45
Sociotechnical Challenges of GenAI
Organized by: Ilias Chalkidis
Location: Salle Les Oudayas Level 2 (stairs)
Description: Discuss the sociotechnical challenges of GenAI development and deployment covering several topics related to the democracy, politics, economics, environment, and more.
Time: 13:30 – 14:15
Queer in AI
Organized by: Sabine Weber
Location: Salle Lamrissa Level 2 (stairs)
Description: Queer in AI’s mission is to raise awareness of queer issues in AI/ML, foster a community of queer researchers and celebrate the work of queer scientists. We use “queer” as an umbrella term for people with diverse non-normative sexual orientations, romantic orientations, and/or genders, corresponding to acronyms like LGBTQIA2S+. We also explicitly include those questioning their identities. We would like to offer an opportunity to meet, network and talk about all things queer in NLP.
Time: 14:30 – 16:00
Masakhane NLP
Organized by: Shamsuddeen Hassan Muhammad
Location: Salle Le Chellah Level 2 (stairs)
Description: BoF Session: Open Data Collection Playbook for AfricaNLP
We are hosting a Birds-of-a-Feather (BoF) session to introduce and co-design the Open Data Collection Playbook for AfricaNLP, a practical, community-driven guide for building high-quality datasets for training and evaluating language models in African languages.
The session will share early outputs from our recent grant from Masakhane and invite participants to contribute lessons learned, early feedbacks, tools, and reusable templates across the full dataset lifecycle: data sourcing and collection, annotation design and quality control, data cleaning and documentation, ethical review and consent, and responsible release practices (e.g., licensing, datasheets, governance). We will also discuss challenges that are especially salient in African contexts, multilingual and multi-dialect settings, code-switching, low-resource annotation pipelines, and community-centered approaches, and identify priority modules for the first public release of the playbook.
Participants will leave with a clear view of the playbook’s scope, ways to contribute, and a shared roadmap toward open, reproducible, and locally grounded dataset creation practices for AfricaNLP.
More info: https://www.masakhane.io/masakhane-african-languages-hub/projects
Mechanistic Interpretability vs. Neuro-symbolic Reasoning in LLMs
Organized by: Fabio Massimo Zanzotto; Leonardo Ranaldi
Location: Salle Les Oudayas Level 2 (stairs)
Description: Are you at all struck by the remarkable knack LLMs have for untangling complex reasoning tasks? Do you have a particular keenness for cracking the code of their inner workings? If so, you’ve likely found yourself delving into the realms of mechanistic interpretability or neuro-symbolic reasoning.
In that case, this is the very BoF for you! After all, observing a model’s properties and truly grasping its internal machinery are simply two sides of the same coin.
Prof Fabio Massimo Zanzotto, Dr Andre Freitas and Dr Marco Valentino will be invited to take part in the discussion. This idea stems from a previous meeting held during the 3rd MathNLP hosted at EMNLP 2025.
Multilingual and Low-Resource NLP: Challenges and Lessons from Practice
Organized by: Soumedhik Bharati
Location: Salle Lamrissa Level 2 (stairs)
Description: This shared interest meeting brings together researchers and practitioners working on multilingual and low-resource NLP. The session will focus on practical challenges encountered when building, training, and evaluating NLP systems for underrepresented languages, including data scarcity, transfer failures, evaluation gaps, domain mismatch, and deployment constraints. Topics may include multilingual pretraining, cross-lingual transfer, low-resource fine-tuning, retrieval and RAG in multilingual settings, and lessons learned from negative or unexpected results. The meeting is discussion-driven and open to participants at all career stages.
Friday, March 27, 2026
Time: 9:30 - 10:30
ACL Mentorship: Navigating Mental Health in Fast-Paced AI Research
Organized by: Oana Ignat, Irene Strauss, Zhijing Jin
Location: Salle Lamrissa Level 2 (stairs)
Description: The ACL Mentorship program creates a space where students and early-career researchers can ask question directly to experienced researchers in NLP and AI. Many students today face intense pressure - from rapid advances in AI, competitive publishing expectations, uncertainty around jobs, and common struggles such as impostor syndrome, perfectionism, anxiety, and stress. This panel will open a conversation about how researchers navigate these challenges while building meaningful careers.
Time: 11:00 – 12:30
Tokenization & Beyond
Organized by: Ilker Kesen
Location: Salle Les Oudayas Level 2 (stairs)
Description: This event aims to gather researchers exploring tokenization and beyond, from developing alternatives to BPE and analyzing tokenizer behavior to tokenization-free language processing approaches that operate on characters, bytes, and visually rendered text. The session will feature lightning talks, themed discussions, and opportunities to identify open problems and future directions.