Workshops

To learn more about each workshop, including schedule & times, visit the workshop website via hyperlink on each workshop’s name.

Wednesday, October 28th

10th Workshop on Automated Knowledge Base Construction (AKBC)

Time: 9:00–17:30
Room: TBA

Description: While Large Language Models (LLMs) have revolutionized NLP, they remain prone to hallucinations, reasoning “mode-collapse” in open-ended generation, and a lack of factual provenance. The Automated Knowledge Base Construction (AKBC) workshop addresses a key missing piece of the generative era: structured knowledge. Knowledge Bases (KBs) serve as ground truth for fact verification, the semantic backbone for constrained decoding in generation, and as a resource behind Retrieval-Augmented Generation (RAG). As such, they are essential tools for mitigating hallucinations and preserving long-tail knowledge. Bringing back AKBC in 2026 is timely. First launched in 2010, and held as an independent conference from 2020-2022, the workshop took a brief hiatus. However, the rapid rise of generative AI, the field’s renewed interest in neuro-symbolic methods, and the growing need for verification in high-stakes domains, has made the integration of structured knowledge more critical than ever. AKBC at EMNLP 2026 will provide a much-needed venue for NLP researchers tackling the intersection of symbolic knowledge and generative models.

3rd Workshop on Natural Language Processing meets Climate Change (ClimateNLP)

Time: 9:00–17:30
Room: TBA

Description: In the past decades, the importance of addressing climate change has been fully adopted by the scientific community. However, complicated dynamics between policy, regulation, industry, and society make it difficult to encourage effective climate action. In recent years, NLP approaches have received much attention as a potential remedy for such complexities. In the 1st and 2nd ClimateNLP workshops, we have learned that there is a continuous need for providing a platform for different stakeholders at the intersection of NLP and climate change. These include not only NLP researchers, but also policymakers, regulators, non-profit organizations, firms, and the general public. The 3rd edition of the ClimateNLP workshop seeks to continue to provide such a platform, aiming to bring together NLP researchers and other stakeholders, and shed light on the unique challenges, opportunities, and potential methods for applying NLP techniques in helping to combat climate change.

11th Workshop on Financial Technology and Natural Language Processing (FinNLP - ACL SIG-FinTech)

Time: 9:00–17:30
Room: TBA

Description: The FinNLP workshop aims to explore the intersection of Natural Language Processing (NLP), Machine Learning (ML), and Large Language Models (LLMs) within the financial, economics, and legal domains. The goal is to foster interdisciplinary research and innovation, addressing the multifaceted challenges inherent in these fields. FinNLP is also the annual workshop of the ACL SIG on Financial and Economic Natural Language Processing (SIG-FinTech). Since its inception in 2019, FinNLP has served as a pivotal workshop dedicated to advancing NLP applications in the financial technology domain. By collocating with major conferences such as EMNLP, IJCAI, IJCNLP-AACL, LREC, and COLING, FinNLP has successfully bridged the AI and NLP research communities, with all proceedings openly accessible via the ACL Anthology. Looking ahead to 2026, the workshop will join forces with multiple teams active in the FinAI and FinNLP communities to curate a richer and more comprehensive program, fostering deeper exchanges and cross-disciplinary collaboration at the intersection of financial artificial intelligence and natural language processing.

Improving Language Models through Learning from Dialogue Game Interaction (LM Playschool)

Time: 9:00–17:30
Room: TBA

Description: The LM Playschool Workshop aims to bring together researchers interested in language learning from social interaction in conversational, collaborative, task-oriented, multi-turn environments. Relevant research strands include: (1) language games as a means of evaluating large language models and their ability to use language in interaction; (2) connections between human language acquisition and machine language learning, with a particular focus on the role of social interaction in data-efficient learning; (3) game-based and interactive environments for training large language models.

4th Workshop on Mathematical Natural Language Processing (MathNLP)

Time: 9:00–17:30
Room: TBA

Description: The articulation of mathematical arguments is a fundamental part of scientific reasoning and communication. Across many disciplines, expressing relations and interdependencies between quantities is at the centre of scientific argumentation. Nevertheless, despite its importance, the application of contemporary NLP models for inference over mathematical text remains under-explored or subject to important limitations. MathNLP represents a forum for discussing new ideas to advance research on Mathematical Natural Language Processing, welcoming novel contributions on model architectures, evaluation methods and downstream applications.

6th Workshop on Multilingual Representation Learning (associated with SIGTYP) (MRL)

Time: 9:00–17:30
Room: TBA

Description: Despite the importance of research on multilingual NLP topics, there are limited venues for this work. In 2025, across NAACL, ACL, EMNLP, and AACL, the MRL workshop was the only workshop focusing on multilingual NLP broadly, without a focus on a particular region or set of languages. MRL in part serves to connect language- and region-specific multilingual work, encouraging crosslingual collaboration and knowledge sharing. This sixth edition brings together an inter-disciplinary community of researchers working on different aspects of multilingual representation learning methods. The main objectives are: (1) to present a wide array of multilingual representation learning methods, including their theoretical formulation and practical aspects; (2) to provide a better understanding of how language typology may impact the applicability of these methods and motivate the development of more generic or competitive methods across languages; (3) to promote collaborations in developing novel software libraries and benchmarks for implementing or evaluating multilingual models that would accelerate progress in the field.

Time: 9:00–17:30
Room: TBA

Description: The NLLP workshop series explores methods and applications of Natural Language Processing (NLP) for the Legal Domain by focusing on original research on legal data and data with legal relevance. On the one hand, legal text (e.g., case law, legal rules, contracts) has distinct characteristics such as specialized vocabulary, particularly formal syntax, and domain-specific semantics, to the extent that legal language is often classified as a sublanguage (i.e. legalese), which makes it challenging for generic NLP tools—even large pre-trained models—to work accurately. On the other hand, the Internet is full of text with legal significance (e.g. detecting advertising language, dark patterns, etc.), as national and supranational regulators pave the way to a new market focused on public interest technology such as consumer forensics.

5th Workshop on NLP for Positive Impact (NLP4PI)

Time: 9:00–17:30
Room: TBA

Description: The increasing integration of Natural Language Processing (NLP) technologies and systems into daily life opens up various opportunities to drive positive social impact. While much of existing research has focused on detecting and mitigating harm—such as hate speech detection and mitigating misinformation—there is a growing need to explore how NLP can address broader societal challenges. Our workshop aims to fill this gap by encouraging more creative application of NLP in support of the UN Sustainable Development Goals, with applications ranging from healthcare and education to tackling climate change, poverty, and inequality. To achieve this potential, we invite interdisciplinary experts across diverse domains to explore how NLP can be effectively applied for social good.

Speech and Audio Language Models Workshop (2nd Edition) (SALMA)

Time: 9:00–17:30
Room: TBA

Description: SALMA (Speech and Audio Language Models) workshop investigates speech and audio as first-class modalities for large language models (LLMs), addressing challenges that are underexplored in mainstream LLM research. Unlike text, audio presents continuous temporal structure, multi-scale semantics, paralinguistic cues, interaction dynamics, and unique safety risks such as deepfakes and impersonation. SALMA focuses on large audio-language models (LALMs) that unify speech, non-speech sounds, and music, exploring architectural design, representation learning, training paradigms, data creation, alignment, and evaluation. Building on the inaugural SALMA workshop at ICASSP 2025, this edition seeks to highlight advances from single-turn post-training to audio-native foundation models capable of multi-turn dialogue, long-context reasoning, multilingual interaction, and controllable generation. The workshop will feature peer-reviewed papers, invited talks, a panel discussion, and a shared task using MMAU-Pro to benchmark audio understanding.

11th Workshop on Natural User-generated Text (W-NUT)

Time: 9:00–17:30
Room: TBA

Description: The 11th Workshop on Natural User-generated Text (W-NUT) addresses the evolving challenges at the intersection of noisy, user-generated written content and modern Large Language Models (LLMs). As LLMs are increasingly deployed in real-world applications, they must contend with the messy, informal, and diverse nature of human communication found on social media, forums, messaging platforms, and other user-generated sources. Recent research shows that while LLMs demonstrate impressive capabilities across many NLP tasks, their robustness to noisy and unstructured text remains a critical challenge. This workshop bridges traditional NLP research on noisy text with contemporary challenges in building, evaluating, and deploying robust LLMs that can handle the complexities of natural language as it actually appears in the wild.

Language Understanding in the Human-Machine Era (LUHME)

Time: 9:00–17:30
Room: TBA

Description: Large language models (LLMs) have revolutionized the way interactional artificial intelligence (AI) systems are developed by making them accessible to the general public. However, such systems are still prone to brittleness in language understanding, which raises doubts about the extent to which they can truly understand human language(s). The “Language Understanding in the Human-Machine Era” (LUHME) workshop aims to reignite, retrieve, resume, and refocus the enduring debate about the role of understanding in natural language use and related applications. It will convene researchers interested in the intersection of language understanding and the effective use of language technologies in human-machine interaction.

Tenth Widening NLP Workshop (WiNLP)

Time: 9:00–17:30
Room: TBA

Description: Widening NLP is a long-running diversity and inclusion workshop series within the ACL community. WiNLP exists because many barriers to participation are upstream of paper quality: unequal access to mentorship and professional networks, reduced opportunities for visibility and senior feedback, financial and visa constraints, and uneven exposure to recruitment and collaboration opportunities. WiNLP addresses these barriers by providing a dedicated, supportive venue for researchers from underrepresented groups (URGs) to present their work, receive feedback, and build lasting professional connections. While presentation opportunities prioritize URGs, allies are encouraged to attend, learn, mentor, and support presenters.

Wednesday, October 28th and Thursday, October 29th

Fourth Arabic Natural Language Processing Conference (ArabicNLP)

Time: 9:00–17:30
Room: TBA

Description: The Fourth Arabic Natural Language Processing Conference (ArabicNLP 2026) is organized by the ACL Special Interest Group on Arabic NLP (SIGARAB). ArabicNLP 2026 follows the First ArabicNLP conference (in 2023 with EMNLP), Second ArabicNLP conference (in 2024 with ACL), Third ArabicNLP conference (in 2025 with EMNLP), and seven previous workshop editions (as WANLP). The research focus of ArabicNLP is Arabic, a collection of language varieties, from Classical to Modern Standard Arabic (MSA), and including many living and historical Arabic dialects. Arabic poses many challenges for computational linguistics, including rich morphology, orthographic ambiguity, and a wide variety of understudied dialects. With over 473 million native speakers and its status as an official UN language, Arabic is an attractive object of study with significant geopolitical relevance. We invite submissions on Core Arabic NLP, Machine Learning for Arabic (including Large Language Models), Applications, and Resources. ArabicNLP 2026 is expected to be a hybrid event featuring shared tasks and a keynote address by Nizar Habash (New York University Abu Dhabi).

Eleventh Conference on Machine Translation (WMT)

Time: 9:00–17:30
Room: TBA

Description: Building on the success of 10 years of Workshops on Statistical Machine Translation (WMT), plus a further 10 years as the “Conference on Machine Translation”, we propose another two-day conference on the topic.

Thursday, October 29th

Accelerating language modeling research with cognitively plausible datasets (BabyLM)

Time: 9:00–17:30
Room: TBA

Description: The BabyLM Workshop challenges the bigger-is-better trend in language model (LM) pretraining by providing a venue for the community to focus on small-scale LM training and exploring its connections to human language development. The workshop will invite research related to data-efficient LMs, developmentally plausible pretraining, small-budget model development, and evaluation issues in small LMs. Following community comments, the 2026 workshop will emphasize multilingual and cross-linguistic data efficiency, and explicitly welcome efficient training studies without a cognitive flavor to encourage cross-fertilization. It will also host the fourth iteration of the BabyLM Challenge, which challenges the community to train LMs on ≤100M words.

Analyzing and Interpreting Neural Networks for NLP (BlackboxNLP)

Time: 9:00–17:30
Room: TBA

Description: As neural networks, and particularly large language models, have become ubiquitous in NLP research and applications, understanding their behaviors and how they connect to their internal mechanisms has evolved from an academic curiosity to a practical necessity. These models now power critical applications yet remain fundamentally opaque: we struggle to predict their failures, understand their reasoning, or ensure their safe deployment. The BlackboxNLP workshop, now in its ninth edition, aims to bring together researchers from machine learning, linguistics, psychology, and neuroscience to develop principled methods for interpreting and analyzing these systems.

Workshop on Document Intelligence and Understanding (DocInsights)

Time: 9:00–17:30
Room: TBA

Description: DocInsights focuses on advancing document understanding beyond plain text by centering structured, semi-structured, and multimodal documents in NLP research. Real-world documents in domains such as law, finance, healthcare, science, and government interleave free-form language with tables, forms, charts, figures, and layout-driven cues, where critical meaning is distributed across elements and pages. While recent document foundation models and multimodal LLMs have improved layout-aware and instruction-following capabilities, substantial challenges remain in structure-sensitive reasoning, grounding and faithfulness to document evidence, robustness across templates and domains, and evaluation that reflects document-specific failure modes under noisy conditions such as OCR or layout errors. Topics span structure-aware representations, multimodal and cross-document reasoning, knowledge integration, robustness and evaluation, and interactive, human-in-the-loop systems for verification and decision support. DocInsights brings together researchers from NLP, information retrieval, vision–language modeling, knowledge representation, and HCI to share models, datasets, benchmarks, and evaluation protocols.

Grounding Language Models: Learning Faithfully and Efficiently (GLM)

Time: 9:00–17:30
Room: TBA

Description: This workshop focuses on grounding language models by connecting them to signals beyond text—such as external knowledge, interaction and feedback, perception, and action—to improve both faithfulness (reducing hallucinations, improving verifiability) and learning efficiency. It brings together researchers from NLP, vision, robotics, and cognitive science to study efficient grounding methods, evaluation protocols, safety, and real-world applications, with an emphasis on post-training, adaptation, and low-resource settings. The goal is to advance language models that learn more reliably and effectively through structured, interactive, and multimodal grounding rather than text-only supervision.

Identifying, Measuring, Preventing, and Assessing Consequences of Bias in Speech LLMs (IMPACT-SPEECH)

Time: 9:00–17:30
Room: TBA

Description: The goal of this workshop is to bring together researchers and practitioners from diverse disciplines—from speech processing, natural language processing, machine learning, social sciences, and ethics—to advance the understanding of bias in speech-based large language models (LLMs). By fostering interdisciplinary discussion, the workshop aims to systematically characterize, analyze, and address bias in speech LLMs, with a particular focus on its sources, interactions, mitigation strategies, and real-world consequences.

Seventh Workshop on Insights from Negative Results in NLP (Insights)

Time: 9:00–17:30
Room: TBA

Description: Publication of negative results is difficult in most fields, but in NLP the problem is exacerbated by the near-universal focus on improvements in benchmarks. This situation implicitly discourages hypothesis-driven research, and it turns creation and fine-tuning of NLP models into art rather than science. Furthermore, it increases the time, effort, and carbon emissions spent on developing and tuning models, as the researchers have no opportunity to learn what has already been tried and failed.

This workshop invites both practical and theoretical unexpected or negative results that have important implications for future research, highlight methodological issues with existing approaches, and/or point out pervasive misunderstandings or bad practices. In particular, the most successful NLP models currently rely on different kinds of pretrained meaning representations (from word embeddings to Transformer-based models like BERT and GPT-3). To complement all the success stories, it would be insightful to see where and possibly why they fail. Any NLP tasks are welcome: sequence labeling, question answering, inference, dialogue, machine translation - you name it. A successful negative results paper would contribute one of the following:

  • broadly applicable recommendations for training/fine-tuning, especially if X that didn’t work is something that many practitioners would think reasonable to try, and if the demonstration of X’s failure is accompanied by some explanation/hypothesis;
  • ablation studies of components in previously proposed models, showing that their contributions are different from what was initially reported;
  • datasets or probing tasks showing that previous approaches do not generalize to other domains or language phenomena;
  • trivial baselines that work suspiciously well for a given task/dataset;
  • cross-lingual studies showing that a technique X is only successful for a certain language or language family;
  • experiments on (in)stability of the previousl y published results due to hardware, random initializations, preprocessing pipeline components, etc;
  • theoretical arguments and/or proofs for why X should not be expected to work;
  • demonstration of issues with data processing/collection/annotation pipelines, especially if they are widely used;
  • demonstration of issues with evaluation metrics (e.g. accuracy, F1 or BLEU), which prevent their usage for fair comparison of methods.

The Workshop for Insights from Negative Results invites short papers as well as non-archival abstract submissions for papers published elsewhere (e.g. in one of the main conferences or in non-NLP venues). Our goal is to provide not only a publication venue, but an opportunity to discuss the most urgent methodological issues, and to think about where the field is going.

13th Web-as-Corpus Workshop (WaC-13)

Time: 9:00–17:30
Room: TBA

Description: The 13th Web-as-Corpus workshop provides a multidisciplinary forum for researchers to address the full life-cycle of web-based data. We invite contributions that introduce or evaluate methodologies for the collection, cleaning, and enrichment of web corpora, as well as their downstream applications across various fields. We particularly welcome contributions that address multilingual web data and less-resourced languages, going beyond an English-centric focus. Our scope encompasses four core pillars: (1) the engineering of high-quality multilingual datasets for foundation models, including data extraction, language identification, data enrichment and filtering, and their evaluation; (2) the use of web data for empirical linguistic research; (3) the analysis of web-scale corpora, both in terms of quality and representativeness and to track societal trends and cultural shifts; (4) the discussion of ethical and legal implications of collecting and using web data.

Multimodal Interaction in Face-to-Face Dialogue (MINT)

Time: 9:00–17:30
Room: TBA

Description: MINT: Multimodal Interaction in Face-to-Face Dialogue is a one-day workshop that targets a core limitation in current “multimodal NLP”: most approaches treat nonverbal signals (gesture, facial expression, gaze, body pose, prosody) as auxiliary context to language, or as post-hoc outputs generated after the linguistic content has been decided. Since human communication is fundamentally face-to-face and meaning is jointly constructed through coordinated verbal and nonverbal behaviours, this creates a major gap in how we model and evaluate real interaction. The workshop brings together researchers from NLP, computational linguistics, computer vision, and cognitive science to advance resources, modelling, processing, and evaluation of multimodal behaviours in situated dialogue. MINT will invite both archival and non-archival contributions across computational models, cognitive/linguistic insights, multimodal datasets, evaluation methods, and interactive applications. The program includes invited talks by Judith Holler and Vera Demberg, and is planned as a hybrid event.

Workshop on Open Reasoning Across Cultures & Languages (ORACLE)

Time: 9:00–17:30
Room: TBA

Description: ORACLE is an inaugural full-day workshop aimed at advancing Open Reasoning Across Cultures & Languages, with a focus on making multilingual reasoning in GenAI more transparent, reusable, and reproducible. The workshop addresses the fact that reasoning capabilities remain uneven across languages due to typological variation, cultural mediation, code-switching, and data scarcity, which often leads to degraded performance and unreliable evaluation outside English. ORACLE will bring together research on multilingual training strategies, architectures, evaluation methodologies, and domain-grounded applications (e.g., science, education, finance, healthcare). A central component is a shared task on Open Cultural Reasoning Across Languages (ORACLE–CR), where participants generate both answers and structured reasoning traces under closed-book and retrieval-augmented settings, evaluated for accuracy, grounding, parity, and cultural appropriateness.

Pluralistic AI & NLP: Diversity-aware, Sociotechnical and Responsible Alignment (PANDORA)

Time: 9:00–17:30
Room: TBA

Description: The PANDORA workshop explores methods, applications, and normative discussions around diversity-aware, sociotechnical, and responsible alignment of NLP technologies. We position pluralism as a core sociotechnical design principle and aim to unite researchers from NLP, AI safety, and related fields. Together, we seek to develop robust methods and shared insights that advance human-centered NLP research and shape future directions in responsible AI alignment.

Second Workshop for REsearch on Agent Language Models (REALM)

Time: 9:00–17:30
Room: TBA

Description: Agents with a high degree of autonomy and self-direction are rapidly transitioning from research prototypes to real-world deployments, transforming both industry and academia. Powered by large language models (LLMs) and enhanced with external tools, knowledge sources, planning strategies, and multi-agent collaboration, these agents are becoming increasingly capable of solving complex, open-ended tasks. However, scaling such systems introduces unprecedented challenges in their development, deployment, evaluation, and governance. To address these challenges, the REALM workshop aims to bring together researchers, practitioners, and thought leaders to chart the future of agentic AI. By bridging technical innovation with responsible practice, this workshop seeks to advance the development of agentic AI that is robust, trustworthy, and impactful.

Third Workshop on Uncertainty-Aware NLP (UncertaiNLP)

Time: 9:00–17:30
Room: TBA

Description: Human language is inherently ambiguous and variable, yet much of NLP research assumes that uncertainty can and should be resolved. This workshop provides a forum for research that models and evaluates uncertainty arising from language, data, and modeling choices. We focus on both aleatoric uncertainty, stemming from the nature of language and data, and epistemic uncertainty, linked to modeling assumptions and limited supervision. Uncertainty-aware NLP spans model design, data collection, learning, inference, evaluation, and deployment, and is particularly critical for low-resource settings and high-stakes applications where reliable confidence estimation is required. UncertaiNLP welcomes contributions on formal uncertainty representations, data and annotation practices, uncertainty-aware modeling and learning, probabilistic inference, applications such as hallucination detection and selective prediction, and principled evaluation and calibration methods.

10th Workshop on Online Abuse and Harms (WOAH)

Time: 9:00–17:30
Room: TBA

Description: Artificial Intelligence (AI) and digital technologies have transformed how people communicate and interact, creating rich societal benefits. At the same time, they have amplified abusive and harmful content, reinforced inequalities, and increased risks to human rights and well-being. Addressing online harms remains a complex challenge spanning technical, social, legal, and ethical dimensions. The Workshop on Online Abuse and Harms (WOAH) is a leading forum for research on these issues. Its 10th edition, themed “Ten Years of WOAH: Reflecting on Progress and New Frontiers”, reflects on how online harms have evolved—from hate speech and harassment to radicalisation, child sexual exploitation, gender-based abuse, disinformation, and algorithmic bias—and how research and policy responses have adapted. WOAH 2026 aims to assess progress, identify emerging challenges, and set priorities for the next decade. WOAH welcomes submissions from diverse fields, including NLP, ML, computational social science, law, psychology, sociology, and cultural studies, and encourages interdisciplinary work, contributions on under-resourced languages, non-archival submissions, and reports from civil society.

2nd Workshop on Sign Language Processing (WSLP 2026)

Time: 9:00–17:30
Room: TBA

Description: Across the world, more than 430 million people experience disabling hearing loss. For many in the Deaf and Hard of Hearing community, sign languages (SL) are the primary means of communication. SLs are not visual equivalents of spoken languages but full-fledged linguistic systems with their own grammar, structure, and cultural significance. However, a shortage of interpreters, as well as limited technological support, create significant accessibility barriers. Sign language processing (SLP), which includes recognition, translation, and generation, can help to surmount these issues, but the field is faced with unique challenges. Unlike spoken languages, SLs are multimodal, involving hand gestures, facial expressions, and body posture. Combined with the scarcity of large annotated datasets and standardized benchmarks, this makes SLP particularly complex. The 2nd Workshop on Sign Language Processing (WSLP 2026) aims to provide a forum for researchers and practitioners across all areas of SLP. While we welcome submissions on any sign language and task, we particularly encourage work focusing on underrepresented languages and communities, with the goal of broadening the inclusivity and impact of sign language technologies.