The MATHBAC project seeks innovative methods for developing agentic AI, explicitly rejecting research focused solely on incremental advancements. Funded by DARPA, the initiative is structured into two phases. The first phase emphasizes deriving the mathematics of agentic AI to enhance communication between systems. The second phase aims to establish tools for a groundbreaking new science, tackling fundamental scientific and mathematical challenges essential for advancing collective agentic intelligence. DARPA has set an ambitious timeline of 34 months to accomplish these goals and is currently inviting proposals from interested organizations. By fostering transformative research, MATHBAC aims to pioneer advancements in AI technology and collective intelligence. Organizations participating in MATHBAC will contribute to potentially revolutionary developments in the field, driving significant progress beyond traditional methodologies. This initiative underscores the importance of innovative thinking in AI research and the pursuit of complex, interdisciplinary solutions.
Source link
