On Wednesday, May 15, 2024, at 12:28 AM, Matt Mahoney wrote: > The top entry on the large text benchmark, nncp, uses a transformer. It is > closed source but there is a paper describing the algorithm. It doesn't > qualify for the Hutter prize because it takes 3 days to compress 1 GB on a > GPU with 10K cores.
If we have MLP, KAN, NN1, NN2, etc. the discoverer could perhaps use the Principle of Least Action on the mathematical system to find/generate the structure that minimizes bit flips to produce similar results. It would be the laziest structure… or lazier structures since the laziest might not be provable. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T1af6c40307437a26-M790672f3424e5b9a96e27236 Delivery options: https://agi.topicbox.com/groups/agi/subscription
