The 2nd International Workshop on LLM-Driven Agents (LLMA 2026)
From Tool to Partner: The Agentic AI Paradigm Shift
Large Language Models (LLMs) have evolved from powerful text generators into autonomous agents capable of decision-making, tool use, and complex task execution. The rise of LLM-driven agents marks a fundamental paradigm shift in AI, where foundation models serve not just as passive engines but as interactive, goal-directed systems. Despite remarkable progress, significant theoretical and practical challenges remain in realizing this shift from tool to partner. Key questions arise: What formal models can underpin LLM-driven agency? How do we ensure safety, alignment, planning consistency, and grounded memory when AI acts as a partner rather than just an executor? Furthermore, moving agentic AI from controlled lab settings to real-world partnerships across domains, such as software engineering, healthcare, education, and robotics, introduces urgent concerns around robustness, scalable deployment, rigorous evaluation, and cost-efficiency.
This workshop aims to bridge the gap between theory and practice by uniting researchers and practitioners working on various aspects of LLM-driven agents.
- Formal models of agentic behavior in LLMs
- Planning, reasoning, and memory mechanisms in agent systems
- Safety, alignment, and interpretability of LLM agents
- Evaluation benchmarks, protocols, and metrics for agentic performance
- Resource efficiency, optimization, and deployment challenges for LLM agents
- Multi-agent collaboration, coordination, and communication
- Human-AI interaction and collaborative agent systems
- Multi-modal and embodied LLM agents
- Tool-augmented and API-integrated agent architectures
- Real-world agentic applications across domains
All submissions for this workshop must be written in English and conform to the Springer proceedings format with the following page limits: 12 pages for papers including references. Submitted papers will undergo a “double-blind” review process, coordinated by the Program Committee. To ensure anonymity of authorship, authors must ensure that authors’ names, affiliations, funding, and any other identifying information of authorship do not appear on the title page or elsewhere in the paper. Authors must provide the complete and final list of authors at the submission stage. No addition, removal, or change in the order of authors is allowed after submission.
Please use one of the following templates for the LNCS (Lecture Notes in Computer Science) format: https://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines
All paper submissions for this workshop will be via Microsoft CMT (under “The 2nd International Workshop on LLM-Driven Agents”).
The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.
| Paper Submission | June 17, 2026 |
| Acceptance Notification | July 20, 2026 |
| Camera Ready | August 25, 2026 |
*All deadlines are 23:59 Anywhere on Earth time.
- Prof. Yong Zhang, Tsinghua University, [email protected]
- Prof. Chun Zeng, Henan University, [email protected]
- Prof. Guigang Zhang, Institute of Automation (CAS), [email protected]
- Prof. Bohan Li, Nanjing University of Aeronautics and Astronautics, [email protected]
- Prof. Yuxi Zhou, Tianjin University of Technology, [email protected]
Xin Wei (Andrew), [email protected]
