I am interested in discussing AGI, not because I have it all figured out, but because I don't. You started to say: "Simple learning, statistics transcending..." So you think that learning is not simple and statistical-based approaches are totally necessary for AGI? (Maybe "necessary" is too strong a term for you?) I think the programming is very difficult. That should be a natural conclusion to draw based on what I have been saying and doing. So, because the programming problem is so difficult I have to start off with something that is extremely simple. However, I am starting to think now that humanity's earlier impression that the process of learning is fairly straight forward (even though the ability to learn complicated subjects is difficult) might be right. Perhaps the years of clamoring after the footsteps of logical methods, semantic vectors, weighted reasoning, probability methods, Neural Networks and Bayesian Networks have caused you to overlook some fresh tracks in the snow. Yes you are right I should get on with it. But I did recently say that I have been having trouble accepting the unsophisticated methods that I have to use to start with. So your demand that I get going is not timely because it is insensitive. On the other hand, now that I have answered you I have come to see that if my simple learning theory is right I should be able to get something going even though my recognition algorithms are awful. The simple learning theory, if it works, might compensate for a lot of other crude algorithms. And if the mechanisms of the simple learning theory can be used in initial recognition problems, as I assume they can be, then I may be able to improve my recognition algorithms as I work the details of some of my theories out. People can become annoyed when a simple idea that resonates somewhere in the back of their head doesn't quite make sense to them. Is your demand that I figure it out before I talk about an acknowledgement that you wouldn't be able to help me anyway? Maybe you're underestimating your ability to think creatively. Jim Bromer
> Subject: Re: [agi] A Very Simple AGI Project > From: [email protected] > Date: Mon, 29 Jul 2013 19:10:30 +0200 > To: [email protected] > > Trolling aside, and who can resist a little trolling here and there since our > global village lacks not its global village idiots, constructive criticism > takes at least two counterparts, and one cannot pretend to be shocked that > such a meeting of minds only takes place a few times a year, after all this > channel is too noisy for most people. Certainly your ruminations don't fit > very well in a professional AGIer's universe, and I would only call myself a > semi-pro at this stage, so quite a few of us thought ok, get on with it and > maybe we will find something to sink our teeth into later on. I think we've > all agreed to disagree in our intuitions, given the complexity of the problem > it would be natural if only one of us found the right checks and balances, > perhaps by genius or sheer luck, but that kind of has to wait for the proof > of the pudding. Again, as far as I am concerned, we share a lot of concerns, > for AIXI for example, while we can't see eye to eye on others. Only way out: > work on that freaking pudding! Simple learning, statistics transcending, show > me the money! > > AT > > On 29.07.2013, at 02:07, Jim Bromer <[email protected]> wrote: > > > One of the interesting but disappointing things about groups like this is > > that there is pretty serious lack of insightful criticism and comment. > > > ------------------------------------------- > AGI > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/24379807-f5817f28 > Modify Your Subscription: https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com ------------------------------------------- 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
