From: jeffrey E. <[email protected]> Sent: Monday, March 6, 2017 10:50 PM To: Joscha Bach Subject: Re: watch the octupus, find its prey, is it =reative.? is it intelligent? are virus creatvie?=C2* finding new methods to infect the host. . hackers=C240 On Mon, Mar 6, 2017 at 6:47 PM, Joscha Bach wrote: It seems that much of what we call intelligence co=ld be described as detecting causal structure in a domain. Causality means=that you have a model that tells you how change of a condition will influe=ce the distribution of outcomes. Of course, crabs do that too, to an exten=. When I was in Munich, a patent lawyer challenged me =o come up with a definition of creativity from a machine learning perspect=ve. He has to judge the originality of inventions, and something that a do=ain expert can come up with "naturally" does not count as =n invention, even if it has never been done before. Lea=ning systems tend to move along gradients, and the "natural soluti=n" of the domain expert can be seen as an extrapolation of the sol=tion space along the known gradients, until an optimum is found. Creativit= seems to involve crossing discontinuities in the problem space, i.e. do m=re than follow the gradients. It is not magic, of course; you can use evol=tionary methods/Monte Carlo etc., but you find that many domain experts ar= capable without being creative. Can we mea=ure the ability of different animals for creativity? <=iV> Am Mar 6, 2017 um 3:23 PM schri=b Jeffrey E <mailto:[email protected]» wrote: So you don't think th=t practical intelligence is the ability to build arbitrary prosthetics? I agree that humans share an important similarity wi=h cells, i.e. they have mechanisms that suppress individual competition wh=n it conflicts too much with group selection. Without such mechanisms, gro=p selection fails, because within each group, defectors have a local advan=age. The simplest mechanism to enforce cooperatio= is reputation, but it wor