KG4Health

The KG4Health workshop addresses the critical intersection of Knowledge Graphs (KGs) and Artificial Intelligence in the biomedical and healthcare domain. While modern AI models (i.e., Generative AI) have shown immense potential in processing healthcare data, clinical text, and biomedical information issues regarding factual accuracy, explainability, and data siloization persist. Knowledge Graphs provide a structured, interpretable foundation that can ground AI models in established medical truths. We welcome computer scientists, medical informaticians, and clinicians to submit their late-breaking results, demos, or position statements via OpenReview to join this high-intensity, half-day dialogue on the future of evidence-based, explainable medical AI. We invite submissions on themes including, but not limited to:

• KG-LLM Integration: Methods for embedding KGs into LLMs to enhance factual grounding and reasoning in medical applications.
• Neuro-symbolic AI: Combining LLMs with KGs (e.g., Retrieval-Augmented Generation - RAG) for medical QA.
• KG Construction & Enrichment: Automated extraction of medical entities/relations from clinical notes and literature.
• Interoperability: Mapping KGs to standard ontologies (SNOMED-CT, UMLS, FHIR).
• Explainable AI (XAI): Using graph paths to provide the "why" behind AI-driven clinical predictions.
• Medical Applications of KG: Drug repurposing, rare disease diagnosis, and personalized treatment pathways.

Motivation

The rapid adoption of AI in medicine and healthcare has highlighted a fundamental "trust gap." While these models are linguistically proficient, they lack an inherent mechanism for factual verification, which can lead to plausible and sometimes dangerous medical hallucinations. AIME participants are increasingly focused on moving AI from experimental environments to high-stakes clinical settings where precision is non-negotiable. Knowledge Graphs (KG) offer the necessary structural scaffolding to ground these generative models in evidence-based medicine, ensuring that AI-driven insights remain consistent with established biological and clinical truths. Furthermore, healthcare data remains notoriously fragmented across disparate silos—ranging from molecular sequences to Longitudinal Electronic Health Records. As the AIME community prioritizes integrative AI, there is a timely need to discuss how KGs act as a universal semantic layer that harmonizes these diverse data sources. This half-day workshop provides a high-intensity forum to discuss the transition from purely statistical AI to "Structured Reliability," enabling the development of transparent, interoperable, and clinically safe healthcare solutions.

Audience

Researchers in semantic web, medical informatics, data scientists in healthcare, and clinicians interested in explainable AI.

• Computer Scientists & AI Researchers: Those specializing in Knowledge Graphs, Large Language Models (LLMs), Neuro-symbolic AI, and Semantic Web technologies.
• Biomedical Informaticians: Researchers focused on ontology engineering, data integration (FHIR, HL7), and clinical natural language processing.
• Clinicians & Medical Professionals: Healthcare practitioners interested in the safety, explainability, and reliability of AI-driven decision support tools at the bedside.
• Health System Leaders & Policy Makers: Individuals tasked with the governance, ethics, and deployment of robust AI solutions within hospital infrastructures.
• Industry Innovators: Data scientists from the pharmaceutical and health-tech sectors working on drug repurposing and evidence-based AI applications.
• Graduate Students (Learners): Doctoral and Masters students at the intersection of AI and Medicine (who may also be interested in the AIME Doctoral Consortium).

Contributions

• Full research papers (6-8 pages).
• Demo, Short, Position or Vision papers (4-6 pages).

Format

• Authors should follow Springer’s guidelines for authors and use their proceedings templates, either for LaTeX or for Word, to prepare their papers. Springer’s proceedings LaTeX templates are also available in Overleaf. Springer encourages authors to include their ORCIDs in their submissions.

Review and Publication

• Please, share your contribution before the deadline through the OpenReview platform.

• All accepted papers (full, short & demo) will be part of the conference proceedings, which will be published by Springer as part of Lecture Notes in Artificial Intelligence (LNAI), a subseries of LNCS dedicated to artificial intelligence research. Papers should be formatted according to Springer’s LNCS format, and one of the categories (Full paper, Short paper, Demo paper) must be selected at the time of submission.
• For a paper to be published, at least one of its authors must be registered. One full registration is required per accepted paper. Please note that there is a strict one paper per registration rule: each paper requires a separate registration. Proceedings from prior AIME conferences can be viewed on SpringerLink.
• Proceedings from prior AIME conferences can be viewed on SpringerLink.
• Authors of the best accepted papers may be invited to expand and refine their manuscripts for possible publication in the Elsevier journal Artificial Intelligence in Medicine.
• Those submitting to the workshop must be aware of OpenReview's moderation policy for newly created profiles:
  • New profiles created without an institutional email will go through a moderation process that can take up to two weeks.
  • New profiles created with an institutional email will be activated automatically.
  • Camera Ready Instructions is available here.

Submit Now

• Fatemeh Bagheri, Saint Mary's University, Canada
• Zhendong Sha, Cleveland Clinic Research, USA
• Ahmet Soylu, Kristiania University College, Norway
• Hannah Kim, Temple University, USA
• Akhil Chaudhary, Toronto Metropolitan University, Canada
• Faezeh Ensan, Cape Breton University, Canada
• Christophe Debruyne, University of Liège, Belgium
• Majid Ziaratban, Saint Mary's University, Canada
• Jaber Rad, Dalhousie University, Canada
• Allard Oelen, German National Library of Science and Technology, Germany
• Hande Küçük McGinty, Kansas State University, Manhattan, KS, USA
• Zaynab Mousavian, Emory University, USA

For information about the conference venue, please refer to the AIME2026 website.