The 3rd International Workshop on Causality, Agents and Large Models (CALM-26)

In conjunction with The 17th International Conference on Ambient Systems, Networks and Technologies

April 14-16, 2026, Istanbul, Türkiye.

Description
Causality, Agents and Large Models (CALM) represents three rapidly-growing fields within artificial intelligence research. This workshop aims to establish methodological integration between these disciplines by investigating how causal reasoning, agentic AI, and explainable AI (XAI) techniques can enhance transparency, decision-making, and adaptation in multi-agent systems (MAS). Agentic AI, which emphasizes autonomous goal-driven behavior and self-reflective reasoning in AI systems, introduces new paradigms for control, coordination, and accountability in intelligent agents. The workshop will provide a forum for researchers to discuss theoretical foundations, practical applications, and future directions at the intersection of causal AI, agentic AI, XAI, Large Language Models (LLM), and MAS.

Workshop goals
Explore the role of causal reasoning and agentic AI in enhancing decision-making, coordination, and adaptation in multi-agent systems.
Discuss technical challenges and opportunities for integrating causal AI, agentic AI techniques, and LLM into MAS frameworks.
Develop methodologies and metrics to evaluate the explainability and autonomy of causal reasoning in agents and large models.
Foster interdisciplinary collaboration between researchers in causal reasoning, XAI, agentic AI, LLM, and MAS.
Examine how explanations support user trust, cognitive ergonomics, and effective human-agent collaboration.
Identify promising directions for future research and development in this rapidly advancing research domain.

Topics
The main topics of the CALM-25 workshop are (but not restricted to):

Theoretical foundations of causal reasoning in Multi-Agent Systems (MAS)
Causal reasoning capabilities in Large Models (e.g., Large Language Models, LLMs)
Theoretical and practical Agentic-AI for autonomous and reflective agents
Explainable AI (XAI) with active inference techniques
Integration of causal and agentic reasoning for adaptive multi-agent coordination
Applications of XAI techniques in agent-based modeling and simulation
Human-centered evaluation of causal explanations in MAS, LLMs, and agentic systems
Causal inference in complex and dynamic multi-agent environments
Mechanistic Interpretability of Causality in Large Language Models
LLMs for coordination, cooperation, and communication among agents
Challenges and opportunities for incorporating XAI and agentic principles into MAS frameworks
Case studies and empirical evaluations of XAI approaches in agents
Generative AI as preprocessing for MAS
Exploring causality with deep generative models
Digital twins and simulators for interpretable synthetic data generation
Graph neural network causal learning
Interpretable and ergonomically-grounded root cause analysis methods for agent decision-making
Logic and argumentation-based approaches to causal reasoning
Ethical and Responsible XAI and Agentic AI in LLM
Human Factors in XAI and Agentic AI
Adaptive and Personalized Explanations (Context-aware, and Human-centric)
Multi-modal explanations and Cross-cultural ergonomics
Ergonomic evaluation of explanation modalities (visual, textual, interactive)
Explainability in human-agent teaming: Ergonomic principles for effective collaboration in mixed teams (humans + AI agents)

Important Dates
Submission deadline: December 20th, 2025
Notification: January 20th, 2026
Final date for camera-ready: January 28th, 2026
Workshop: April 14-16, 2026
Submission
Submitted papers must be no longer than 6 pages including all figures, tables and references. The submitted paper must be formatted according to the Guidelines of Procedia Computer Science, Elsevier. Authors are required to write the **The 3rd International Workshop on Causality, Agents and Large Models (CALM-26)** in the title section of the template, and not the name of the main conference. Kindly use CALM-26 Template

Attention: Do not modify the templates supplied by Elsevier in any way (font size, margins, …). Elsevier will not publish the papers before *all* manuscripts do comply with the original templates. Any deviation from the templates in a manuscript will delay the publication for all papers.

Submissions will be reviewed by 2-3 members of the program committee, who are experts in the field. The acceptance of the submitted papers will be based on scientific rigor, methodological soundness, contribution to causality and explainability in agents and large models, and originality.

Authors are requested to submit their papers electronically as PDF files via the EasyChair submission page: EasyChair

Proceedings
All accepted papers will be scheduled for oral presentations. At least one author of each accepted paper is required to register and attend the conference to present the work. **Camera ready files** submission webpage can be found here (Will be active by November 30th, 2025)

All accepted and registered papers will be published by Elsevier Science in the open-access Procedia Computer Science series on-line. Procedia Computer Science is hosted by Elsevier and on Elsevier content platform ScienceDirect, and will be freely available worldwide. All papers in Procedia will be indexed by Scopus and by Thomson Reuters’ Conference Proceeding Citation Index. All papers in Procedia will also be indexed by Scopus and Engineering Village (Ei). This includes EI Compendex. Moreover, all accepted papers will be indexed in DBLP. The papers will contain linked references, XML versions and citable DOI numbers.

Workshop Chairs
Dr. Yazan Mualla (Belfort-Montbeliard University of Technology, France). yazan.mualla[at]utbm.fr
Dr. Hui Zhao (Tongji University, China). huizhao[at]tongji.edu.cn
Dr. Amro Najjar (Luxembourg Institute of Science and Technology, Luxembourg). amro.najjar[at]list.lu

Advisory Board
Prof. Dr. Stéphane Galland (Belfort-Montbeliard University of Technology, France)
Prof. Dr. Abdeljalil Abbas-Turki (Belfort-Montbeliard University of Technology, France)

Program Committee
TBA

Registration
Please visit the Registration for more information.

Venue
Please visit the Venue for more information.

Previous Iterations
CALM-2024, Proceedings (Springer): Advances in Explainability, Agents, and Large Language Models.

CALM-2025, Proceedings (Springer): TBP.