Jim,
thanks. I was thinking about how we use prediction for survival. Without prediction I would put my hand in the fire and leave it there, because I would not be able to predict that fire causes pain. Or that food is good for hunger. Just like a tree. Locomotion goes with prediction, without it I would be able to avoid pain, or seek food. Just like a tree. That's why we have a brain, to predict and to move. Sergio From: Jim Bromer [mailto:[email protected]] Sent: Wednesday, June 20, 2012 2:53 PM To: AGI Subject: Re: [agi] Prediction Did Not Work (except in narrow ai.) On Wed, Jun 20, 2012 at 9:01 AM, Sergio Pissanetzky <[email protected]> wrote: > Jim, > > I see prediction as essential for survival. In order to survive we need to predict the consequences of what we do to the world. We predict by establishing chains of causality, hence the importance of causality in survival. > > You are discussing a different angle. You are discussing the use of prediction for verification of theories. But isn't that also a form of prediction? You predict something, then compare your prediction with what actually happened in the world, and adjust your process of prediction in order to account for what really happened. > > Would you explain your take on this? > > Sergio You seem to be saying that the verification of theories is a form of prediction. Of course people are able to make and utilize predictions. However, there is no strong evidence that this method can be used reliably to produce Artificial General Intelligence without a lot of complications. Even Francis Bacon in discussing his modern view about scientific method, recognized that during a discussion of the observations of a scientific experiment people might disagree on the nature of some of the objects of the observation. He had a rather simple response for this problem. If there was any dispute about an object (or method) used in the observation that object could also be subject to scientific method. The problem with this is that contemporary AGI programs tend to fail in a variety of simple situations. It is easy to talk about adjusting a theory based on comparison amongst human beings but when you are talking about AGI the whole process often fails just because the abilty to recognize what occurred is a major part of the problem. If we were able to create a good general AI program then all these things would be feasible because a general intelligence is able to have an idea about the things that happen around it. If the ground was as simple as a cartoon where the ground is painted with a solid color and any object is painted with a different color then our AI programs could pick out the objects against the ground. But when you are talking about an image provided from a camera of a landscape, slightly different kinds of things may not be of different color and under different conditions a particular color does not always look the same. The real world, or a representation of the variety of events that can occur in the real world is just too complicated and varied to make the problem of strong AI easy. Jim Bromer AGI | <https://www.listbox.com/member/archive/303/=now> Archives <https://www.listbox.com/member/archive/rss/303/18883996-f0d58d57> | <https://www.listbox.com/member/?&; ad2> Modify Your Subscription <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-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
