A0179
Title: Counting cycles with DeepSeek
Authors: Tracy Ke - Harvard University (United States) [presenting]
Abstract: Despite recent progress, AI still struggles with advanced mathematics. A difficult open problem is considered: How to derive a computationally efficient equivalent form (CEEF) for the cycle count statistic? The CEEF problem does not have known general solutions and requires delicate combinatorics and tedious calculations. Such a task is hard to accomplish by humans, but it is an ideal example where AI can be very helpful. The problem is solved by combining the proposed novel approach and the powerful coding skills of AI. Results use delicate graph theory and contain new formulas for general cases that have not been discovered before. It is found that, while AI is unable to solve the problem all by itself, it is able to solve it if a clear strategy, step-by-step guidance, and carefully written prompts are provided. For simplicity, the focus is on DeepSeek-R1, but other AI approaches can also be investigated.
