X-TAIL26: 2nd Workshop on the extraction and exploitation of long-TAIL Knowledge with LLMs and Knowledge Graphs EKAW 2026 Turin, Italy, September 29-October 1, 2026

LLMs encode world knowledge through pre-training on massive datasets, making them the backbone of knowledge extraction tasks. Their reliability degrades on long-tail knowledge: low-popularity knowledge that occurs infrequently in pre-training data. Popularity is not a neutral property: pre-training datasets are predominantly web-crawled and, as such, are generalist, English-centric, and mostly produced over the past 30 years by Western, High-income, Educated, Liberal, Male-dominated (WHELM) communities, raising the risk of models underperforming on specialized domains, non-English languages and non-contemporary times sources, and on knowledge belonging to marginalized social groups. Retrieval-Augmented Generation has been proposed as a mitigation, but corpora used for retrieval may still be biased. Knowledge Graphs (KGs) provide a more transparent and deterministic alternative, yet open-domain KGs such as Wikidata exhibit coverage gaps along the same dimensions.

The X-TAIL workshop aims to advance research on extracting, exploiting, and ultimately preserving long-tail knowledge. It welcomes contributions on: methods to extract knowledge from domain-specific, multilingual, historical, and low-resource language sources, including approaches combining LLMs and KGs; studies on the head/tail knowledge distinction; investigations on how popularity distributes across the specificity, linguistic, temporal, and cultural dimensions of knowledge, and its effect on system per-
formance; characterization of gaps in knowledge bases and mitigation strategies

Submission Guidelines

All papers must be original and not simultaneously submitted to another journal or conference.

• Submission site: EasyChair
• Format: PDF, using the CEUR-WS (CEURART) template
• Page limits (references excluded):
  • Full research papers: 8–12 pages
  • In-Use and Experience papers: 8–12 pages
  • Short research papers: 4–6 pages
  • System/Demo/Position papers: 4–6 pages
• Repositories: Authors are strongly encouraged to submit accompanying repositories (e.g., code, data, and models) in accordance with FAIR principles. Such materials should be made available via GitHub or a comparable publicly accessible platform.
• Template: CEUR-WS Overleaf template (CEURART)

List of Topics

Topics of interest include (but not limited to):

• Head and tail knowledge definition: operationalisation and computation of knowledge popularity
• Knowledge extraction (Relation Extraction, Entity Linking, KG generation, Question Answering) from long-tail sources (domain-specific, non-English, non-contemporary, related to marginalised social groups);
• Context-augmented methods (RAG and KAG) optimised for long-tail knowledge
• Knowledge graphs generation and completion for mitigating coverage gaps in existing knowledge bases;
• Computational methods for long-tail knowledge processing in specialised domains: health, law, finance, sustainability, digital humanities (computational literary studies, computational history, etc.)
• Impact of multilingualism on model performance on long-tail knowledge
• Long-tail knowledge representation in ontologies and KGs;
• Benchmarks that systematically address specificity, linguistic, temporal, and WHELM biases of both LLMs and KGs
• Error analysis of system performance stratified by knowledge popularity and its intersection with domain, language, social group, and time dimensions;
• Negative results on methods developed to mitigate underperformance on long-tail knowledge;
• Studies of harms perpetuated by popularity-driven knowledge hierarchies learned by LLMs and reflected in KGs

Publication

X-TAIL-26 proceedings will be published in CEUR Workshop Proceedings.

Organizing Committee

• Lia Draetta - University of Turin, Italy
• Arianna Graciotti - University of Groningen, The Netherlands
• Enrico Daga - The Open University, UK
• Aidan Hogan - DCC, Universidad de Chile, Chile

Venue

Co-located with EKAW 2026 the 25th International Conference on Knowledge Engineering and Knowledge Management, September 29th - October 1st, 2026, University of Torino, Italy

Contact

All questions about submissions should be emailed to lia.draetta@unito.it