Important Dates
| Milestone | Date (AoE, UTC−12) |
|---|---|
| Abstract submission | January 28, 2026 |
| Paper submission | February 4, 2026 |
| Author notification | March 19, 2026 |
| Camera-ready submission | April 23, 2026 |
| Workshop | May 25–26, 2026 |
Topics
Over decades, substantial knowledge and expertise have been developed in the engineering of Multi-Agent Systems (MAS), including theories, architectures, languages, platforms, and methodologies for designing, implementing, and deploying autonomous agents and multi-agent systems. Despite this strong foundation, the emergence and integration of augmented language models and generative agents into agentic and multi-agent systems, introduces new challenges and opportunities for MAS engineering to evolve and address this expanding landscape of autonomous agents.
In this context, EMAS 2026, with its special theme on Hybrid Agent Architectures and Multi-Agent Systems, welcomes contributions that revisit, extend, or challenge established MAS engineering approaches, explore emerging generative agent architectures and agentic systems, or examine the integration of different approaches into hybrid agent architectures and multi-agent systems. Submissions may address and extend foundational questions in MAS engineering, including but not limited to:
- How to specify, design, implement, verify, and test multi-agent systems in light of emerging technologies such as generative agent models, agentic systems, and neuro-symbolic Artificial Intelligence (AI), as well as modern application needs including explainability, interoperability, and flexible tool use.
- Which (multi-)agent architectures and languages are best suited to meet diverse design objectives and system requirements.
- How elements from established and emerging agent architectures can be combined to engineer hybrid architectures that leverage strengths while mitigating weaknesses of individual approaches.
- How established MAS principles can inform the design of generative and hybrid agent architectures.
- How advances in agentic technologies can drive the re-examination and extension of MAS principles, models, and languages, addressing emerging forms of generative agent interaction, coordination, and reasoning.
- How multi-agent systems can enable interoperability among agents of heterogeneous architectures while ensuring effective, governed coordination and collaboration.
- How to engineer agents and multi-agent systems that are verifiable, explainable, transparent, and accountable by design.
- Which processes and methodologies can integrate the above to provide a disciplined approach to MAS engineering.
EMAS 2026 provides a forum for researchers and practitioners in agent-oriented software engineering, multi-agent system programming, declarative agent languages and technologies, generative agents, and agentic systems to present and discuss research and emerging results in MAS engineering. The workshop aims to:
- Enhance knowledge of the theory and practice of engineering autonomous agents and multi-agent systems, advancing the state of the art.
- Demonstrate how MAS methodologies, architectures, languages, and tooling can be applied in the engineering of hybrid multi-agent systems and agents built on hybrid architectures.
- Define new directions for MAS engineering by combining established MAS approaches with insights from the Agentic AI community.
- Encourage PhD and Master’s students to engage with and contribute to the field.
Submissions
We solicit four types of submissions:
- Regular papers should: (1) clearly describe innovative and original research; or (2) report a survey on a research topic in the field; or (3) explain how existing techniques have been applied to a real-world case. (16 pages, excluding references, in LNCS format).
- Short papers should: (1) describe novel and promising ideas and/or techniques that are in an early stage of development; or (2) present a vision for some part of the field, including challenges, and research opportunities (see the AAMAS Blue Sky Track CFP for more information on these sort of papers). (8 pages, excluding references, in LNCS format).
- Student papers should describe M.Sc. or Ph.D. research in the field of engineering multi-agent systems. The paper should clearly describe the problem tackled and why it is important, the research method, the (expected) contributions of the research, and the evaluation. The lead author on the paper should be the student. (6 pages, excluding references, in LNCS format).
- Tools, testbeds, and demo papers should describe a novel tool or demonstration in the field of engineering multi-agent systems. Submissions may range from early prototypes to in-house or pre-commercialised products. Authors of other EMAS 2026 papers are also welcome to submit an accompanying tool/demo paper. The paper should provide a link to supplementary material that allows the reviewers to evaluate the submission such as website or video (4 pages, excluding references, in LNCS format).
Submission policy: all papers should be original and not be submitted elsewhere. The review process is single blind: submissions should not be blind, reviewers will be.
Submissions should be formatted following the LNCS formatting style which is available via: http://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines
Abstract submissions are now open via OpenReview: https://openreview.net/group?id=ifaamas.org/AAMAS/2026/Workshop/EMAS
Post-proceedings
Papers accepted to the workshop may be considered for inclusion in the post-proceedings.
To support discussion of emerging ideas, provide opportunities for improvement, ensure a high-quality proceedings volume, and offer authors appropriate publication pathways, a two-stage process is employed: based on the reviewers’ assessments, some papers may be recommended for inclusion in the post-proceedings in their current form, while others may be invited to submit revised and extended versions. Subject to final arrangements, the post-proceedings are intended for publication in Springer’s Lecture Notes in Artificial Intelligence (LNAI) series.
Awards
We will select a paper for the best paper award based on reviewers’ scores and recommendations.