Steve, I am not finished with the summary. It should take me a least a week to finish, maybe two. In the first line of Part 1 I mentioned that complexity is a major problem so when you declare that I am making some rookie errors because I fail to appreciate the speed issue I can only conclude that you either did not read the summary too well or that you are exaggerating. This lessens the magnitude of your criticism, and it actually represents a weak (an extremely weak) positive indicator that maybe I am on the right track. In other words, there may be something subtle in my theories that you are missing. I will be glad to discuss this with you after I am finished with the summary. Jim Bromer
On Sat, Apr 13, 2013 at 4:30 PM, Steve Richfield <[email protected]>wrote: > Jim, > > You are making some rookie errors... > > 1. You fail to appreciate the speed issue. Computers are WAY too slow to > even be able to experiment in the domains you are speculating. In short, > you are at least decades too early to start. Note (for example) my fast > parsing, where I FINALLY proposed a fast-enough method of parsing English, > using ideally-constructed tables, and NOT using the sort of expensive > learning you are talking about. > > People hear "gigahertz" and their eyes cross, their knees weaken, and they > think they can do ANYTHING. > > WATSON comes a little closer, still understands nothing, but uses 2,880 > processors to do it. > > 2. There is a belief/condition in people's minds that they can > arbitrarily discard entire dimensions, often more than one dimension, and > still make a working learning system. I might bet a week of my time testing > such a highly questionable presumption, but certainly not years. In any > case, computers are still too slow for your approach, even with discarded > dimensions. > > Note what I did with my Scanning UV Fluorescent Microscope. Here is > something I first came up with ~50 years ago - and it was WAY ahead of its > time. As late as ~2 years ago it was rejected as being "off topic" by the > AGI conference. Now, Obama is calling for just such a machine in his BRAIN > Initiative. I am now scrambling to get my SUVFM considered because it IS > the best of the several competing approaches. > > I suspect that you may end up doing the same. Once we know how brains > work, and you can buy a petascale machine from Best Buy for ~$1K, then you > can dust off your proposal and forge on ahead. You will then be government > funded (via Social Security) and have your medical insurance covered (by > Medicare) as I now am. > > Mine is a "success" story, as most good designs that are ahead of their > time end up lost to history, often because their creators have also been > lost to history (died). Once you have finished your design, your next job > will be to stay alive for another ~50 years, to be around to promote it > when the "missing pieces" have become readily available. > > Steve > ================== > On Sat, Apr 13, 2013 at 3:39 AM, Jim Bromer <[email protected]> wrote: > >> Part 1 >> >> I feel that complexity is a major problem facing contemporary AGI. It >> is true, that for most human reasoning we do not need to figure out >> complicated problems precisely in order to take the first steps toward >> competency but so far AGI has not been able to get very far beyond the >> narrow-AI barrier. >> >> I am going to start with a text-based AGI program. I agree that more >> kinds of IO modalities would make an effective AGI program better. However, >> I am not aware of any evidence that sensory-based AGI or multi-modal >> sensory based AGI or robotic based AGI has been able to achieve something >> greater than other efforts. The core of AGI is not going to be found in the >> peripherals. And it is clear that starting with complicated IO >> accessories would make AGI programming more difficult. It seems obvious >> that IO is necessary for AI/AGI and this abstraction is a probably more >> appropriate basis for the requirements of AGI. >> >> My AGI program is going to be based on discreet references. I feel that >> the argument that only neural networks are able to learn or are able to >> incorporate different kinds of data objects into an associative field is >> not accurate. I do, however, feel that more attention needs to be paid to >> concept integration. And I think that many of us recognize that a good >> AGI model is going to create an internal reference model that is a kind of >> network. The discreet reference model more easily allows the program to >> retain the components of an agglomeration in a way in which the traditional >> neural network does not. This means that it is more likely that the >> parts of an associative agglomeration can be detected. On the other >> hand, since the program will develop its own internal data objects, these >> might be formed in such a way so that the original parts might be difficult >> to detect. With a more conscious effort to better understand concept >> integration I think that the discreet conceptual network model will prove >> itself fairly easily. >> >> I am going to use weighted reasoning and probability but only to a >> limited extent. >> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >> <https://www.listbox.com/member/archive/rss/303/10443978-6f4c28ac> | >> Modify <https://www.listbox.com/member/?&;> Your Subscription >> <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> > <https://www.listbox.com/member/archive/rss/303/10561250-470149cf> | > Modify<https://www.listbox.com/member/?&;>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-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
