Summary of Key Points
Young mathematician Levent Alpöge used the AI tool Fable 5 to unexpectedly find a counterexample to the Jacobi Conjecture, a century-old problem in algebraic geometry, during the World Cup finals, proving that the conjecture is incorrect. This counterexample is so simple that anyone with knowledge of calculus can verify it, causing a sensation in the mathematical community. It not only ended the futile efforts of many mathematicians but also led to the rejection of related conjectures such as the Dixmier Conjecture. Additionally, the event has drawn significant attention to the potential of AI in mathematical research, with some even speculating that AI could win the Fields Medal in the future. The incident also shed light on Zhang Yitang’s connection to this conjecture and clarified several misconceptions about his doctoral thesis.
I. The Jacobi Conjecture: A Century-Old Mathematical Dilemma
The Jacobi Conjecture is considered a “star problem” in algebraic geometry, which can be simply stated as follows: If the Jacobian determinant of a polynomial map (a measure of whether the function is locally invertible) is a non-zero constant, is the map globally invertible?
Its allure lies in:
- Simple starting point: In high school calculus, if the derivative of a one-dimensional function is always positive, the function is strictly monotonic and therefore invertible. However, this intuition fails in multi-dimensional or complex spaces (for example, the derivative of the exponential function is non-zero but periodic). Mathematicians want to know whether using “the simplest polynomial map” with a Jacobian determinant that remains constant would restore invertibility.
- Significant status: In 1998, Fields Medalist Smale listed it as Problem 16 in the “Millennium Problems,” suggesting that solving it could advance algebraic geometry significantly.
- History of mistaken proofs: Numerous top mathematicians, including the Italian school and algebraic geometer Gröbner, have published purported proofs, but they were quickly identified as flawed—making it one of the most dangerous “pitfalls” in mathematics.
II. An Unexpected Discovery While Watching the Cup
Alpöge’s discovery was more like a sudden “eureka moment”:
1. Inspiration: The World Cup finals were boring, so he discussed the Jacobi Conjecture with friends and casually fed the problem to the AI tool Fable 5.
2. Result: A few hours later, Fable 5 provided a counterexample: a three-dimensional polynomial map whose Jacobian determinant was always -2 (a non-zero constant), yet there were three different inputs (A=(0,0,-1/4), B=(1,-3/2,13/2), C=(-1,3/2,13/2) that all mapped to the same output (-1/4,0,0).
3. Verification: Alpöge confirmed the counterexample using Wolfram Alpha, proving that the conjecture was incorrect. The conjecture claims that if the Jacobian determinant is constant, the map is globally invertible, but this map is not even a bijection, let alone invertible.
The most remarkable aspect is that anyone with basic calculus knowledge can understand and verify this counterexample without advanced algebraic geometry skills.
III. The Ripple Effects: The Counterexample Shatters a Century-Old Conjecture
The impact of this counterexample went beyond the conjecture itself:
- End of futile efforts: Many mathematicians had spent years working on this problem, only to have their work invalidated.
- Rejection of related conjectures: The Jacobi Conjecture and the Dixmier Conjecture (about Weyl algebra automorphisms) as well as the Poisson Conjecture are “stably equivalent”; if one is wrong, so are the others.
- AI’s breakthrough: Fields Medalist Gowers praised this as the first time an LLM (Large Language Model) had solved a major problem he had long admired. Some boldly predict that AI could win the Fields Medal in the future (Gowers also thinks it’s possible in the next few decades).
IV. Zhang Yitang and the Jacobi Conjecture
The incident also brought to light some aspects of Zhang Yitang’s career, clearing up several misconceptions:
- Truth about his thesis: It was rumored that Zhang’s doctoral thesis proved the Jacobi Conjecture to be false, but the existing archive shows it focused on studying the number of field extensions in two dimensions, not directly proving the conjecture. His thesis was approved, and he received his Ph.D. from Purdue in 1991.
- Reasons for career setbacks: His advisor, Mo Zongjian, did not write a recommendation letter for him, preventing Zhang from finding an academic position, forcing him to work part-time. However, the two-dimensional Jacobi Conjecture might still be true (Mo Zongjian had proven that the number of extensions in two dimensions is ≤100 in 1983).
- Clarification of a similar problem: There was a domestic claim of solving the Jacobi Conjecture, but it pertained to an ODE (Ordinary Differential Equation) problem and was unrelated to the algebraic geometry conjecture discussed here.
V. AI’s Entry into the Mathematical Realm: Will It Replace Human Mathematicians?
This event has raised questions about the role of AI in mathematical research:
- AI as a tool, not a replacement: Alpöge is a top mathematician (a Harvard researcher with a Fields Medalist advisor), and AI found the counterexample under his guidance—this was a collaborative effort between human and AI.
- Changing research methods: AI can quickly search through literature and explore numerous possibilities, helping humans avoid mental biases. For example, the counterexample was simple, but human mathematicians might not have discovered it due to their reliance on traditional theories (such as the Implicit Function Theorem).
- Future trends: AI will become an important aid in mathematical research, but human judgment and creativity will remain essential. After all, formulating problems, guiding AI, and verifying results all require a deep understanding of mathematics.
Conclusion
The disproving of the Jacobi Conjecture is not only a major breakthrough in mathematics but also a demonstration of the combination of AI and human intelligence. It shows that while AI is breaking traditional research boundaries, human curiosity and expertise remain the driving force behind scientific progress. In the future, the collaboration between AI and human mathematicians may bring even more unexpected discoveries in the world of mathematics.