Foretelling the Future of AI through Mathematical Foundations - From Learning Architectures to Trustworthy Agents
AAAI 2027 2nd MATH4AI Workshop
February 22-23, 2027 | Montréal, Canada | Palais des congrès
AAAI 2027 2nd MATH4AI Workshop
February 22-23, 2027 | Montréal, Canada | Palais des congrès
Topics
We encourage the paper and poster submission on the following topics and hope to encourage the discussion on the mathematical foundation in these topic in the community:
Approximation Theorem of new architecture
Theoretical analysis of the expressivity and approximation capabilities of emerging learning architectures, including foundation models, neural operators and other structured models.
Mathematics-inspired optimization algorithms
Optimization and training algorithms motivated by mathematical principles, with emphasis on convergence, training dynamics, generalization, efficiency, and stability.
New Architectures Inspired by Mathematical Theories
Design and analysis of learning architectures based on ideas from geometry, symmetry, operator theory, dynamical systems, numerical analysis, and scientific computing. Examples include physics-informed models, neural operators, and structure-preserving architectures.
Mathematical Foundations of AI Agents.
Mathematical frameworks for representing and understanding reasoning, planning, memory, uncertainty, tool use, feedback, and sequential decision-making in AI agents.
Stability and Reliability of Agentic AI Systems.
Mathematical analysis of agent trajectories, including convergence, robustness, verification, uncertainty, and error propagation across multi-step reasoning and decision processes.
Submission requirements
We accept both poster/short papers with no more than 4 pages and long papers with no more than 7 pages. The page number does not include the references and appendix. All accepted papers will be presented as posters. The program committee will select a few papers for oral presentation.
The submissions should adhere to the AAAI paper guidelines. The contributions will be non-archival but will be hosted on our workshop website. All contributions will be peer reviewed (single-blind).
For any questions email us at math4ai.nvaitc@gmail.com
Submission Site Information
Papers should be submitted via OpenReview
Important Dates
Submission Deadline: November 20, 2026
Notification of Acceptance: December 2, 2026