>MATT MAHONEY=====> My design would use most of the Internet (10^9 P2P nodes). ED PORTER=====> That's ambitious. Easier said than done unless you have a Google, Microsoft, or mass popular movement backing you.
>> ED PORTER=====> I mean, what would motivate the average American, or even the average computer geek turn over part of his computer to it?... >MATT MAHONEY=====> The value is the ability to post messages that can be found by search, without having to create a website. Information has negative value; people will trade CPU resources for the ability to advertise. ED PORTER=====>It sounds theoretically possible. But actually making it happen in a world with so much competition for mind and machine share might be quite difficult. Again it is something that would probably require a major force of the type I listed above to make it happen. >> ED PORTER=====> Are you saying that as a system becomes bigger it naturally becomes unstable, or what? >MATT MAHONEY=====> When a system's Lyapunov exponent (or its discrete approximation) becomes positive, it becomes unmaintainable. This is solved by reducing its interconnectivity. For example, in software we use scope, data abstraction, packages, protocols, etc. to reduce the degree to which one part of the program can affect another. This allows us to build larger programs. In a message passing network, the critical parameter is the ratio of messages out to messages in. The ratio cannot exceed 1 on average. ED PORTER=====> Thanks for the info. By "unmaintainable" what do you mean? I don't understand why more messages coming in than going out creates a problem, unless most of what nodes do is relay message, which is not what they do in my system. The unruly chaotic side of AGI is not something I have thought much about. I have tried to design my system to largely avoid it. So this is something I don't know much about, although I have thought about net congestion a fair amount which can be very dynamic, and that sounds like it is a related to what you are talking about. I have tried to design my system as a largly asynchronous messaging system so most processes are relatively loosely linked, as browsers and servers generally are on the internet. As such, the major type of instability I have worried about is that of network traffic congestion, such as if all of a sudden many nodes want to talk to the same node, both for computer nodes and pattern nodes. I WOULD BE INTERESTED IN ANY THOUGHTS ON THE OTHER TYPES OF DYNAMIC INSTABILITIES A HIERARCHICAL MEMORY SYSTEM -- WITH PROBABILISTIC INDEX-BASED SPREADING ACTIVATION -- MIGHT HAVE. Matt, it sounds as if OpenCog ever tries to build a large P2P network "you the man". Ed Porter -----Original Message----- From: Matt Mahoney [mailto:[EMAIL PROTECTED] Sent: Tuesday, December 04, 2007 7:42 PM To: [email protected] Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] --- Ed Porter <[EMAIL PROTECTED]> wrote: > Matt, > > IN my Mon 12/3/2007 8:17 PM post to John Rose from which your are probably > quoting below I discussed the bandwidth issues. I am assuming nodes > directly talk to each other, which is probably overly optimistic, but still > are limited by the fact that each node can only receive somewhere roughly > around 100 128 byte messages a second. Unless you have a really big P2P > system, that just isn't going to give you much bandwidth. If you had 100 > million P2P nodes it would. Thus, a key issue is how many participants is > an AGI-at-Home P2P system going to get. My design would use most of the Internet (10^9 P2P nodes). Messages would be natural language text strings, making no distinction between documents, queries, and responses. Each message would have a header indicating the ID and time stamp of the originator and any intermediate nodes through which the message was routed. A message could also have attached files. Each node would have a cache of messages and its own policy on which messages it decides to keep or discard. The goal of the network is to route messages to other nodes that store messages with matching terms. To route an incoming message x, it matches terms in x to terms in stored messages and sends copies to nodes that appear in those headers, appending its own ID and time stamp to the header of the outgoing copies. It also keeps a copy, so that the receiving nodes knows that they know it has a copy of x (at least temporarily). The network acts as a distributed database with a distributed search function. If X posts a document x and Y posts a query y with matching terms, then the network acts to route x to Y and y to X. > I mean, what would motivate the average American, or even the average > computer geek turn over part of his computer to it? It might not be an easy > sell for more than several hundred or several thousand people, at least > until it could do something cool, like index their videos for them, be a > funny chat bot, or something like that. The value is the ability to post messages that can be found by search, without having to create a website. Information has negative value; people will trade CPU resources for the ability to advertise. > In addition to my last email, I don't understand what your were saying below > about complexity. Are you saying that as a system becomes bigger it > naturally becomes unstable, or what? When a system's Lyapunov exponent (or its discrete approximation) becomes positive, it becomes unmaintainable. This is solved by reducing its interconnectivity. For example, in software we use scope, data abstraction, packages, protocols, etc. to reduce the degree to which one part of the program can affect another. This allows us to build larger programs. In a message passing network, the critical parameter is the ratio of messages out to messages in. The ratio cannot exceed 1 on average. Each node can have its own independent policy of prioritizing messages, but will probably send messages at a nearly constant maximum rate regardless of the input rate. This reaches equilibrium at a ratio of 1, but it would also allow rare but "important" messages to propagate to a large number of nodes. All critically balanced complex systems are subject to rare but significant events, for example software (state changes and failures), evolution (population explosions, plagues, and mass extinctions), and gene regulatory networks (cell differentiation). -- Matt Mahoney, [EMAIL PROTECTED] ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?&; ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244&id_secret=72126189-b12bb2
<<attachment: winmail.dat>>
