[FoRK] [fonc] Deployment by virus

Stephen Williams sdw at lig.net
Thu Jul 19 09:16:36 PDT 2012


Using genetic programming (GP)[1] and spin offs like meta-optimizing semantic evolutionary search (MOSES)[2], you can grow and 
select for useful code.  Programming such a system mainly consists of putting constraints and fitness measures in place and 
providing the system with appropriate building block operations.  Since the building blocks could be used to solve a wide range 
of problems, it is a potentially powerful system for dynamically solving many problems. We'll need a lot more progress, and even 
more CPU/memory/storage/bandwidth, to make this work competitively, but it looks feasible at some point soon.

[1] http://en.wikipedia.org/wiki/Genetic_programming
[2] http://wiki.opencog.org/w/Meta-Optimizing_Semantic_Evolutionary_Search

sdw

On Thu Jul 19 01:43:39 2012, Eugen Leitl wrote:
>
> ----- Forwarded message from John Nilsson <john at milsson.nu> -----
>
> From: John Nilsson <john at milsson.nu>
> Date: Thu, 19 Jul 2012 02:22:50 +0200
> To: Fundamentals of New Computing <fonc at vpri.org>
> Subject: [fonc] Deployment by virus
> Reply-To: Fundamentals of New Computing <fonc at vpri.org>
>
> I just had a weird though, maybe there is some precedence?
>
> If we were to do software development in a more organic manner,
> accepting the nature of complex systems as being... complex. In such a
> setting we might have no blue-print (static source code) to usable for
> instantiating new live systems ex nihilo, or the option to take down
> existing systems to "deploy an upgrade". The code running the nodes
> can be the result of wild mutation or complex generative algorithms.
>
> A mode of development could be to work on prototypes in a lab, a clone
> or an isolated node from the production system. When the desired
> properties are created in the prototype they would then spread through
> the production system by means of a virus which would adapt the new
> properties to the running instances individually according to their
> unique configuration.
>
> Is it feasible? Would it provide new options? Any research done in
> this direction?
>
> BR,
> John
>
> -- 
> Stephen D. Williams sdw at lig.net stephendwilliams at gmail.com LinkedIn: http://sdw.st/in
> V:650-450-UNIX (8649) V:866.SDW.UNIX V:703.371.9362 F:703.995.0407
> AIM:sdw Skype:StephenDWilliams Yahoo:sdwlignet Resume: http://sdw.st/gres
> Personal: http://sdw.st facebook.com/sdwlig twitter.com/scienteer


More information about the FoRK mailing list