Long papers
Original and unpublished work with the same page limit as the EMNLP 2026 main conference; must be anonymized.
NLP/CORE Unranked
WNUT 2026
The 11th Workshop on Natural User-generated Text
W-NUT 2026 is a workshop collocated with EMNLP 2026 that focuses on natural user-generated text, ranging from noisy text processing to real-world applications of large language models (LLMs). It emphasizes challenges in multilingual, low-resource, and informal language contexts, including social media, messaging platforms, and cross-platform variations.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Short papers
Original and unpublished work with the same page limit as the EMNLP 2026 main conference; must be anonymized.
Research areas in scope
01
Classical NLP Tasks on Noisy Text
NLP of noisy text, e.g., POS and NER tagging, parsingText normalization and error correctionParaphrase identification and semantic similarity of short text or noisy textExtracting user demographics, profiles, and major life eventsMachine translation and multilingual NLP over noisy textInformation extraction from noisy text and event extractionColloquial language, e.g., idiom detectionDomain adaptation to user-generated textDetecting rumors, contradictory information, sarcasm, and humor on social mediaSentiment analysisTemporal aspects of user-generated content (resolving time expressions, concept drift, etc.)Representing and mining language variation in user-generated content
02
LLMs and Noisy Text
Robustness of LLMs to noisy, ungrammatical, or informal inputTraining and fine-tuning of LLMs on user-generated textEvaluation methodologies for LLMs on noisy textDomain adaptation of LLMs to specific user-generated text domainsLLM-generated noisy text: detection, characteristics, and implicationsInstruction-following in LLMs when faced with noisy or ambiguous promptsRetrieval-augmented generation (RAG) with noisy documents
03
Real-World LLM Applications
LLM performance on text from social media, forums, and messaging platformsHandling code-switching, slang, and emerging language in LLMsLLMs for content moderation and safety in user-generated contextsMultilingual and cross-lingual LLM applications on informal textLLMs for assisting language learners and processing learner textBias, fairness, and representation in LLMs trained on user-generated text
Policies worth checking twice
- Submissions must be original and unpublished.
- Long and short paper submissions must be anonymized.
- Dual-submission is allowed, but authors must indicate at submission time if the paper has been or will be submitted elsewhere.
- Authors of accepted papers must notify organizers by the camera-ready deadline whether the paper will be presented or withdrawn.
- Submissions must conform to ACL style guidelines.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026