Jim, Again, I think I see the POV to solve this. All animals, from single cells to us, are fundamentally adaptive process control systems. We use our intelligence to live better and more reliably, procreate, etc., much as single-celled animals, only with MUCH richer functionality. Everything fits this hierarchy of function leading to intelligence.
Then, people like those on this forum start by ignoring this and trying to create intelligence from whole cloth. This may be possible, but there is NO existence proof for this, no data to guide the effort, etc. In short, there is NO reason to expect a whole-cloth approach to work anytime during the next century (or two). However, some of the mathematics of adaptive process control is known, and I suspect the rest wouldn't be all that tough - if only SOMEONE were working on it. I suspect that when the answers are known, it will be a bit like spread spectrum communications, where there is a payoff for complexity, but where ultimately there is a substitute for designed-in complexity, e.g. like the pseudo-random operation of spread spectrum systems. Genetics seems to prefer designed-in complexity (like our brains) but there is NO need for computers to have such limitations. Whatever path you take, you must "see a path" to have ANY chance of succeeding. You must have a POV that helps you to "cut the crap" in pursuit of your goal. Others here are working on whole-cloth approaches, yet bristle when challenged for lacking a guiding POV. I see some hope in adaptive control math. Perhaps you see something else, but it MUST have an associated guiding POV for you to have any hope of succeeding - more than a simple list of what it does NOT have. Steve On Fri, May 1, 2015 at 5:20 PM, Jim Bromer <[email protected]> wrote: > The other views that I forgot to add is that the knowledge objects > have to components that can be combined in various ways and that there > are no absolute elementary knowledge object. Every kind of knowledge > object including the particles that more coherent objects are made of > have the potential to be opened, explored and related to the greater > world of concepts (knowledge objects.) > > Jim Bromer > > > On Fri, May 1, 2015 at 7:01 AM, Jim Bromer <[email protected]> wrote: > > How can I describe the features and behaviors of a group of > > hypothetical algorithms that would contain the potential to achieve > > advances in AI so that I have some basis for actually designing them? > > The first step is to describe some algorithms from narrow AI and then > > show that my algorithms should, hypothetically, be stronger than them. > > > > You might just say that stronger AI is going to need all kinds of > > algorithms but that does not give you enough to start thinking about > > how you might design an advanced AI program. > > > > First of all, stronger AI has to be more than learning to associate > > particular responses to particular inputs. It also has to be more than > > mere numerical extrapolation or interpolation based on the use of a > > numerical method that represents some particular problem. > > > > So then simplistic reinforcement, for example, is not - in itself - > > enough. Numerical methods - in themselves - are not going to be > > enough. > > > > One thing that I realized while trying to talk to Mike Tintner was > > that true AI needs varying field sizes in order to hold enough > > variation to avoid degenerating into simplistic extrapolations and > > interpolations. (The data type does not have to be variable as long as > > they can be used in strings and fields which are.) > > > > You need some kind of trial and error in stronger AI. You also need to > > recognize that knowledge objects are not typically commensurate. So > > your program needs to be able to fit the pieces of knowledge together > > to see what makes sense and what does not. It needs to discover what > > might be relevant to some situation and what needs to be 'translated' > > from one knowledge object to another. It needs to recognize that even > > though two or more knowledge objects may be relevant, the features may > > not fit against the relevant features of the others. You might use > > simple association or numerical correlation to designate these poorly > > fitting parts but it is often going to take greater insight to > > effectively integrate the different kinds of relevant knowledge > > objects. > > > > So finally I think that Stronger AI is going to need > > reason-based-reasoning. (I still cannot understand how people in these > > AI discussion groups have actually denied that.) In order to learn how > > to use reasons effectively they will need to be integrated with the > > knowledge objects that they are being used with. > > > > Jim Bromer > > > ------------------------------------------- > AGI > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/10443978-6f4c28ac > Modify Your Subscription: > https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > -- Full employment can be had with the stoke of a pen. Simply institute a six hour workday. That will easily create enough new jobs to bring back full employment. ------------------------------------------- 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
