OpenAI announces three big science claims for researchers
Says it has produced results in mathematics, computer science and physics, and is changing how experts review some of them. The company has published material…
OpenAI announces three big research claims in mathematics, computer science and physics, with outside experts guiding review of the computer science and physics work. The company says it is changing how it releases AI-generated results after a recent batch of mathematical manuscripts.
Two sets of materials are available on GitHub, according to The Verge. OpenAI has not released the full physics material, while researchers and the wider academic community are being asked to assess how such claims should be examined.
Experts guide two of the claims
In mathematics, OpenAI says its work concerns local smoothing for the Euclidean wave equation. The company says the proof uses an auxiliary partial differential equation and establishes Sogge’s local smoothing for waves in every spatial dimension at least two. It also says the result yields the Kakeya maximal and dimension conjectures and Stein’s spherical restriction estimates throughout the interior of the conjectured range.
For this result, OpenAI says it released the model output directly for researchers to work through and develop explanations. That differs from the review process for the other two areas, where external experts are guiding examination of the work.
The computer science paper claims to refute the Exponential Time Hypothesis, a proposition that has been widely believed to be true. OpenAI says the paper presents a deterministic algorithm that exactly counts satisfying assignments of fixed-width CNF formulas in subexponential time and polynomial space. The case of 3-SAT is what the company says refutes the hypothesis.
An independent working group led by Russell Impagliazzo, Ramamohan Paturi and Ryan Williams will examine and explain the proof. OpenAI says the group asked for the materials to be made public immediately so others could inspect them. The company says the group also decided to release the AI output write-up and Lean proof on GitHub.
Physics work remains unpublished in full
OpenAI describes its physics result as a mathematically rigorous existence proof involving certain interacting quantum field theories with gauge fields and fermions in four dimensions. The company says the work includes Quantum Chromodynamics and a unified completion of the Standard Model of particle physics. The full information has not yet been released, The Verge reported.
OpenAI says David Gross will lead a group including Michael Douglas, Arthur Jaffe, Igor Klebanov and Nikita Nekrasov to guide the physics work. The company has not provided the full research materials in the release described by The Verge.
The release process itself has become part of the debate. OpenAI’s panel of mathematics advisers, AGMAI, said it recommended that AI companies not give selected people first access to results. The panel cited unequal access, borrowed reputation and biased selection among its concerns, according to The Verge.
The advisers instead recommended publication through scholarly repositories, alongside Lean-formalized proofs, information about how problems were selected and funding for people who work to understand the results. Those steps would make it easier for researchers to inspect not just a headline claim, but also the supporting work and the path that produced it.
OpenAI also published an open letter seeking community input on early access to AI-generated results, disclosure of conflicts of interest and use of unpublished material in later research. It asked how credit should account for discovery as well as checking, fixing, simplifying, explaining and extending results, and how students and researchers can be supported as research methods change.
The questions carry practical weight for academics: outside review can help test complex claims, while selective early access may leave other researchers without the same opportunity to assess or build on them. The company’s latest release therefore pairs scientific claims with an unresolved discussion about who gets to examine AI-generated work, and when.
Source: theverge.com
