AI 在数学领域的研究取得新进展！ 近日，一家名为 Math, Inc. 的公司宣称利用 Gauss 智能体，已经完成了一个关乎 8 维和 24 维空间中的最优球体堆积定理的形式化证明，代码量约为 20 万行 （LOC）。 这一定理最初由数学家玛丽娜・维亚佐夫斯卡（Maryna Viazovska）及其合作者证明，Maryna Viazovska 也凭此荣获 2022 年国际数学家大会菲尔兹奖。 据悉， 这是本世纪唯一一个被完全形式化证明的菲尔兹奖成果，也是历史上规模最大的单一用途 Lean 形式化项目。 而背后的这家名为 Math, Inc. 的公司，也不是第一次完成数学问题的形式化证明了。 资料显示，Math, Inc. 是由 xAI 前联合创始人、Morph Labs 首席科学家 Christian Szegedy 创立，致力于通过自动形式化技术打造可验证超级智能，Gauss 则是首款 专为协助数学专家开展形式化验证工作打造的自动形式化智能体。 去年， Gauss 只用了三周时间，就完成了陶哲轩和 Alex Kontorovich 提出的数学挑战， 即在 Lean 定理证明器中完成强素数定理（Prime ...

Core Insights - Google has made significant progress with its Gemini model, successfully addressing 13 problems from the Erdős Problems database, including 5 novel solutions and 8 rediscoveries of existing answers [1][2][4]. Research Overview - The Erdős Problems database, named after mathematician Paul Erdős, contains 1,179 problems, with 483 (41%) classified as solved. However, many "open" problems may have existing solutions that were not previously identified [4][5]. - The research utilized a custom AI agent named Aletheia, which employed a natural language verifier to filter approximately 700 open Erdős problems down to 212 potential solutions [9]. Methodology - Aletheia's process involved initial filtering by non-expert mathematicians, reducing candidates to 27, which were then rigorously reviewed by domain experts. Out of about 200 candidates, 137 (68.5%) had fundamental errors, while only 13 (6.5%) provided meaningful answers to Erdős's original questions [9][12]. Key Results - The 13 meaningful solutions were categorized into four types: 1. Autonomous solutions (Erdős-652, Erdős-1051) where Aletheia found the first correct solution, although Erdős-652 was based on existing literature [14]. 2. Partial AI solutions for multi-part problems (Erdős-654, Erdős-935, Erdős-1040) [15]. 3. Independent rediscoveries (Erdős-397, Erdős-659, Erdős-1089) where solutions were already known but not initially recognized [15]. 4. Literature identification (Erdős-333, Erdős-591, Erdős-705, Erdős-992, Erdős-1105) where existing solutions were identified despite being marked as open [15][16]. Research Significance - The findings indicate that AI has reached a level where it can tackle "low-hanging fruit" in mathematical problems, providing a new benchmark for AI research in mathematics. However, the authors caution against overstating the mathematical significance of these results, as they are solvable by any expert in the field [19]. - The study highlights challenges in verifying the originality of solutions and the potential for "unconscious plagiarism" where AI reproduces knowledge from training data without proper citation [19][20].

Core Insights - The article discusses the transformation of AI from a "mathematical problem-solving tool" to a "research collaboration partner," exemplified by Tsinghua University's AI mathematician system (AIM) successfully solving a complex mathematical proof [1][2][3] Group 1: AI's Role in Mathematical Research - The research demonstrates the feasibility of AI as a collaborative partner in tackling complex mathematical problems, marking a significant shift in how mathematical discoveries can be approached [2][3] - The study addresses the limitations of current AI systems in mathematics, which often excel in standardized tasks but struggle with real-world research needs [4][5] - The AIM system's collaboration with human researchers led to a comprehensive 17-page mathematical proof, showcasing the potential of human-AI synergy in advanced mathematical research [8][29] Group 2: Methodological Framework - The research outlines five effective human-AI interaction modes that serve as operational guidelines for AI-assisted mathematical research [13][30] - These modes include Direct Prompting, Theory-Coordinated Application, Interactive Iterative Refinement, Applicability Boundary and Exclusive Domain, and Auxiliary Optimization, each designed to enhance the collaborative process [14][17][19][21][22] - The systematic approach to human-AI collaboration not only improves the efficiency of mathematical proofs but also provides a reusable framework for future research [30] Group 3: Future Directions - The study emphasizes the need for further development of human-AI interaction models to enhance mathematical research capabilities and explore their applicability across different mathematical fields [32][34] - Future research will focus on optimizing the AIM system's architecture to improve its reasoning capabilities and overall performance in mathematical theory research [36]

Core Insights - The collaboration between Terence Tao and GPT-5 Pro successfully addressed a three-year-old unsolved problem in differential geometry, showcasing the potential of AI in academic research [1][10]. Group 1: Problem Solving Process - The original problem involved determining if a smooth topological sphere in three-dimensional space, with principal curvature absolute values not exceeding 1, encloses a volume at least equal to that of a unit sphere [3]. - Tao's initial approach was to restrict the problem to star-shaped regions and utilize integral inequalities, but he sought AI assistance for complex calculations [4]. - GPT-5 Pro completed all calculations in 11 minutes and 18 seconds, providing a complete proof for the star-shaped case using various inequalities, some of which Tao was familiar with, while others were new to him [5]. Group 2: AI's Performance Evaluation - AI demonstrated effectiveness in small-scale problems, contributing useful ideas and only minor errors, but it reinforced Tao's incorrect intuition on medium-scale strategies [11][12]. - In large-scale understanding, AI was beneficial in accelerating research and helping Tao abandon unsuitable methods [14]. - Tao's experience highlighted the necessity of human expertise for further advancements in complex problems, indicating that AI's role is more supportive than substitutive [11][16]. Group 3: Historical Context and Evolution of AI Tools - Tao's exploration of AI's potential in mathematics began with the release of ChatGPT, where initial interactions yielded disappointing results due to a lack of depth in understanding mathematical problems [21][22]. - The introduction of GPT-4 marked a turning point, as it significantly improved efficiency in handling statistical data and mathematical tasks, leading to a more optimistic view of AI's integration into research [22][29]. - Tao's ongoing experiments with AI tools have shown that while AI can assist in numerical searches and problem-solving, it still requires careful oversight to mitigate issues like hallucinations or irrelevant outputs [29][31].