Agentic AI: New Frontiers in Computing, Mathematical Reasoning, and Scientific Discovery
Recent advances in artificial intelligence are transforming not only how we solve computational problems, but also how we formulate research questions, develop algorithms, and discover new knowledge. As AI evolves from a tool for individual tasks into autonomous agents capable of reasoning, coding, experimentation, and verification, we are approaching a fundamental shift in the practice of computational and scientific research. In this talk, I will discuss emerging opportunities and challenges at the intersection of agentic AI, computer science, and applied mathematics. A central theme is the evolving relationship between human and artificial intelligence: while AI can increasingly scale the generation of ideas, algorithms, experiments, and mathematical arguments, scientific judgment, rigorous verification, and trustworthy reasoning remain critical bottlenecks. I will highlight three emerging paradigms—agentic computing, autonomous scientific discovery, and agentic theorem proving—and discuss the new algorithmic and mathematical foundations they demand. I will conclude with a research outlook on trustworthy autonomous agents, efficient AI-driven computation, multi-agent collaboration, and AI-enabled scientific discovery. These directions point toward a future in which AI does more than accelerate existing research workflows: it fundamentally expands what we can compute, reason about, and discover.