Search
OpenAI previously published 722 AI math manuscripts. Later, due to a sign error in a key paper involving the Hodge Conjecture, three related manuscripts were withdrawn, 14 were revised, and the citations of 13 were updated, bringing the total number of manuscripts to 719. This withdrawal only targeted the proof process; the related propositions themselves were not deemed incorrect. The original manuscripts remain accessible, and the core conclusion scope of the original Hodge Conjecture-related findings has also been adjusted. Currently, approximately 42% (300 papers) of the main results in the manuscripts have been formalized using Lean, and nearly all the withdrawn and substantially revised content comes from sections that have not undergone Lean formalization.
-
C114 Communication Network -
Communication Home
