I posted before about the history of AI, where neural nets were invented in the 1950s and nearly killed in the 1970s and 1980s in favor a logical symbolic programming. The logical approach hit a wall around 2000, and around 2010 neural nets started making rapid progress, leading to the AI LLM.
What is not so well known is that the logical symbolic programming came back, and got integrated into the AILLMs. Something like the Riemann Hypothesis could never have been proved with just neural nets. The use of hard logic models was essential.
Update: The proof was verified.
This comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteDeleted the comments, but have cached them (with this post) to a PDF. Can make them available if the interest is bona fide. -- Ajit
DeleteSeems to be a non-constructive proof. REALLY proving RH would produce an explicit function of the Prime Counting Function with an error term established by Helge Von Koch. They don't have that.
ReplyDelete