The landscape of mathematics is being forever changed for better or worse with the advent of state-of-the-art artificial intelligence (AI) tools. Gauss is at the center of this revolution. Created by Math, Inc., it combines exploratory natural language logical reasoning with complete formalized logical reasoning. This innovative approach allows mathematicians to enhance their understanding of complex mathematical concepts and verify intricate proofs with remarkable efficiency.
Jesse Han started Math, Inc. to change the way that mathematical proofs are verified. They’ve set Gauss up as an industry-leading Alternative Procurement tool. The AI’s contributions are evident in notable projects, including the automation of Viazovska’s proof for sphere packing in 24 dimensions. This partnership between human mathematicians and AI technology demonstrates one of the powerful ways that technology can profoundly assist the mathematical community. Combined, they are creating extraordinary new opportunities.
Breakthroughs in Sphere Packing Proofs
In 2016, Maryna Viazovska made waves in the mathematical world by solving the sphere packing problem in two specific dimensions. Her research into the 24-dimensional sphere packing earned her the highest honor—the Fields Medal—award in July 2022. The demand for consistency and readily verifiable proofs, especially for more complicated proofs, has increased since then, attracting the interest of AI solutions.
Recognising the need to build on this work, in March 2024 the Formalising Sphere Packing in Lean project was launched to allow continued exploration of this space. Gauss was key to making this happen. He autoformalized Viazovska’s 24-dimensional proof—the thing has more than 200,000 lines of code!—in two weeks. This notable accomplishment showcases Gauss’s capacity to process large bodies of mathematical manuscript at an accelerated speed and with precision.
Gauss’s capabilities extend beyond the 24-dimensional realm. Most recently, it further autoformalized the 8-dimensional sphere packing proof and accurately detected a typo in the resulting published paper. This correction is an important signal of Gauss’ reliability as a verification tool, and a significant step forward in formally proving the validity of proofs.
“They told us that they had finished 30 ‘sorrys’, which meant that they proved 30 intermediate facts that we wanted proved,” – Hariharan
The Role of AI in Collaborative Efforts
The partnership between AI and human mathematicians magnifies the ability to formalize proofs. This piece by Jesse Han portrays Gauss as an extraordinary reasoning agent. It does so in a truly impressive manner by combining natural language processing with formal reasoning techniques.
“It’s a particular kind of language model called a reasoning agent that’s meant to interleave both traditional natural language reasoning and fully formalized reasoning,” – Jesse Han
On this blog inspiring mathematicians share their ideas in natural language. Yet they lean on Gauss to faithfully and correctly formalize those concepts. The result is a powerful new collaboration that combines the human touch of creativity with the speed and efficiency of machines.
Then, in early January, Han announced a breakthrough that produced a solidly stronger Gauss. It is a significant step that will enable better cooperation between mathematicians and AI, making it possible to explore more sophisticated proofs together.
“We made a research breakthrough sometime mid-January that produced a much stronger version of Gauss,” – Jesse Han
Together, these improvements greatly enhance verification processes. They inspire mathematicians to tackle deep problems that seemed out of reach.
Future Implications and Potential
The progress achieved by Math, Inc. and Gauss heralds an exciting future for mathematical research and proof verification. As AI technologies advance, their further inclusion into the mathematical world opens up thrilling opportunities for new discoveries. Tools like Gauss stand to fundamentally change how mathematicians approach the work at hand. Most importantly, they can make the process of verifying complex proofs scalable and help identify mistakes in published work more easily.
Liam Fowl, a prominent figure in the field, expressed optimism about these new developments:
“These new results seem very, very impressive, and definitely signal some rapid progress in this direction,” – Liam Fowl
Fowl emphasized the importance of formal verification as a means to ensure accuracy in mathematical work:
“Formal verification of a proof is like a rubber stamp,” – Liam Fowl
Hariharan, a first-year PhD student at Carnegie Mellon University who has worked together with Gauss, agrees with this assessment. He describes the excitement surrounding this technology and its potential impact on the field of mathematics:
“But at the end of the day, this is technology that we’re very excited about because it has the capability to do great things and to assist mathematicians in remarkable ways,” – Hariharan
Math, Inc. is currently continuing to improve Gauss and expand its features and functionalities. The mathematical community looks forward to seeing how these improvements will inform and enhance future research projects.