Yeah I've been saying for a while that system identification of PDEs is likely where it's at given the need to hook up with the empirical world's priors with a formal system that's Turing Complete -- at least potentially. Indeed the second connectionist summer would likely not have happened if not for the fact that Charlie Smith was focused on system dynamics as a result of his work as a founder of the Energy Information Administration just prior to being given control of the System Development Foundation's endowment that he used to raise connectionism from the grave to which Minsky and Papert had consigned it.
Which brings up a question that's been bugging me for a while: Who killed the second connectionist summer? I know it had something to do with support vector machines, but there must have been a group/movment in particular that managed to pull off that grug-brained stunt. On Tue, Apr 20, 2021 at 8:40 AM Bill Hibbard via AGI <[email protected]> wrote: > Interesting article: > > https://www.quantamagazine.org/new-neural-networks-solve-hardest-equations-faster-than-ever-20210419/ > > Points to a couple arxiv papers: > https://arxiv.org/abs/1910.03193 > https://arxiv.org/abs/2010.08895 ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T2ce9391dcef2fcc9-Madb9c3dfda18af00fa6e06b1 Delivery options: https://agi.topicbox.com/groups/agi/subscription
