YKY, I like the direction this is going. Is Genifer 3.0 implemented yet? A couple of thoughts: 1. Another way to eliminate variables is to employ analogy (i.e., substitution). Hofstadter's group has been exploring analogy for years. 2. What is the purpose of the registers? 3. Another way to constrain search is to employ activation, (i.e., only search active entries in the knowledge base). That is the principle behind the Neural Propositions in PAM.P2. 4. Now that you've eliminated variables you can begin to do some incredible reasoning via analogy, deduction, and induction: generalization, specialization, abstraction, concretion, similization, dissimilization, etc. Read Michalski's Inferential Theory of Learning. Cheers, ~PM
Date: Wed, 20 May 2015 23:36:40 +0800 Subject: Re: [agi] Genifer 3.0 white paper From: [email protected] To: [email protected] On Tue, May 19, 2015 at 11:37 PM, Matt Mahoney <[email protected]> wrote: What are your goals? What do you hope that Genifer4 will accomplish? Do you have any test cases? What is your cost estimate? What I have in mind is a search engine that can answer questions directly, and return answers that are not just direct matches but also logical consequences of the queries with background knowledge. I have not estimated the development costs... I don't even have a team now :] AGI | Archives | Modify Your Subscription ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
