On 11/10/06, Ben Goertzel <[EMAIL PROTECTED]> wrote:
> The word "agent" is famously polysemous in computer science. In my
> prior post, I used it in the sense of "software agent" not "autonomous
> mental agent." These Novamente MindAgents are just software objects
> with certain functionalities, that get scheduled and exchange
> messages. The same sense of agents used in Gul Agha's work on agent
> architectures, years ago.
> The word "agent" is famously polysemous in computer science. In my
> prior post, I used it in the sense of "software agent" not "autonomous
> mental agent." These Novamente MindAgents are just software objects
> with certain functionalities, that get scheduled and exchange
> messages. The same sense of agents used in Gul Agha's work on agent
> architectures, years ago.
OK, I looked up the dictionary and you're right - "agent" can mean any thing that produce certain effects. It seems that your approach involves using many such agents as in evolutionary algorithm (selection among many individuals) or swarm intelligence. Is that so? If yes then my approach is drastically different; I'd use only 1 or a few agents for NL processing.
YKY
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/?list_id=303
