IT Home October 27 news: According to a report by Chao News on October 26, Fields Medal winner and University of Chicago mathematics professor Deng Yu delivered a keynote speech at last weekend's "Zhishifenzi" (The Intellectual) "Mr. Science" forum, systematically expounding his thoughts on mathematical research in the AI era.
Deng Yu stated that AI's advantages over humans lie, on the one hand, in AI's unlimited breadth of knowledge, and on the other hand, in AI's extremely low cost of experimentation / trial and error. But we must hold firm to one point: whether in mathematical research or in anything else,People always exist as subjects; AI is a tool, a tool that we humans use to conduct research.
Deng Yu believes that if AI is treated as a researcher, it should adhere to academic norms. If AI proves a very major problem, a very reasonable requirement,is to publish the AI's chain of thought. However, AI companies refuse to publish them, which clearly does not conform to academic convention. "This AI company is unwilling to do so. Among all the proofs they have published, only a very small number have actually published the chain of thought; in most cases, theyunder the pretext of so-called anti-distillation, refuse to publish the chain of thought and related information."
Deng Yu stated that for AI companies,you cannot distill boldly when you are the one distilling the web literature of all humanity and the knowledge of all humanity; in fact, no one wants to distill you — people just want to understand the chain of thought, and then you assume others want to distill you and start saying this won't do and that won't do.
Deng Yu proposed that, by open-sourcing mathematical models,ensuring the mathematics community has models of its own in its own hands, there would be no need to depend on such AI companies. In addition, the use of models would follow academic norms, including data usage, publication of the chain of thought, traceability, and so on.
IT之家 noted that on September 11 local time, Terence Tao, Deng Yu, Peter Scholze and 25 Fields Medal winners in total, as initial signatories, jointly issued a statement titled "A Severe Misalignment of AI in Mathematics", warning that a "severe misalignment" is emerging between AI companies and the mathematical community.
The statement said that over the past few months, the mathematical capabilities of large language models have improved significantly, enabling them to solve many major open problems across mathematical fields. However, the push by AI companies to use mathematical problems as solved benchmarks is harmful to the mathematical sciences and the mathematics community.The goals of AI companies are severely misaligned with those of the mathematics community, and these are part of broader alignment issues affecting other sciences and creative professions, and even society as a whole.
The statement mentioned that these AI solutions are often released hastily, without time for formal write-up, nor time to isolate new methods and new ideas, or to cite others' related work. Like all creative professions,this raises serious questions of attribution and plagiarism. Moreover, without mathematicians willing to take responsibility for their development and integration into the mathematical canon, ideas conceived by AI can never be fully realized, and the crucial chain of human transmission among mathematicians will be lost.