On Tue, Nov 25, 2008 at 11:31 AM, Matt Mahoney <[EMAIL PROTECTED]> wrote: > One of the problems in defining RSI in a mathematically vigorous way is > coming up with a definition that is also useful. If a system has input, then > there is really no definition that distinguishes self improvement from > learning, at least not one that people can agree on.
Why would you want to? Who said RSI wasn't about learning? It's entirely about learning! > Of course, a practical AGI is going to have input, so my definition seems to > be of little practical use. Nevertheless there are proposals along these > lines. My goal is to prove the limitations of these systems. Which proposals? By who? Maybe you should cite them in your paper. > One example of a system without input would be a chess playing program that > improved its game by playing itself. One could imagine many approaches. For > example, suppose the program makes random variations in its source code and > plays these copies against each other in timed matches, keeping only the > winning variations. What are the limitations of this approach? Or consider a > more general approach to intelligence, where the parent gives its offspring > hard problems. Is it possible for superhuman intelligence to arise > spontaneously? What enforces the rules of the game? That said, I don't think a self-playing chess program can learn anything useful about chess.. but hey, maybe someone else has said, in print, that they do and made the argument that this is a good way to go about making an intelligent system.. you need to track down that work and address it directly. Otherwise you're just arguing with yourself, and isn't your paper saying that is futile? :) > What I show is that if we measure intelligence by computational efficiency, > then yes, but if we measure it by amount of knowledge, then no. Sure, but I *think* everyone already knew that. If you want to say someone is wrong, you need to say who you are talking to. > Anyway, I appreciate any comments that can be used to improve the paper. AGI > is a hard subject to write about, given the wide range of opinions and the > lack of proven results. On that note, the paragraph on batch vs interactive is not indented the same as the rest of the paragraphs. Trent ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=120640061-aded06 Powered by Listbox: http://www.listbox.com
