[P2P-F] Fwd: Automenta : How to Develop Automatic Intelligence (AI)

Tue Feb 8 06:08:02 CET 2011

---------- Forwarded message ----------
From: Dante-Gabryell Monson <dante.monson at gmail.com>
Date: Tue, Feb 8, 2011 at 8:51 AM
Subject: Fwd: Automenta : How to Develop Automatic Intelligence (AI)
To: econowmix at googlegroups.com, sustainable_solidarity at yahoogroups.com,
hc_ecology at yahoogroups.com

http://blog.automenta.com/2011/02/how-to-develop-automatic-intelligence.html

<http://2.bp.blogspot.com/_mWGkAr1Xo5Y/TVCK9OiJX7I/AAAAAAAAAzU/MHgvCcX19sM/s1600/automenta.png>

Consider and improve the human-computer interaction ergonomics of all
software tools involved in development, in order to result in a completely
integrated code/data editor and visualizer, rendering the entire system
reflectively transparent to both humans and software agents within itself
for maximally productive development processes.

<http://3.bp.blogspot.com/_mWGkAr1Xo5Y/TVCLuCW67cI/AAAAAAAAAzY/PEXoDjiCvVM/s1600/Ergonomic-Office-Personal-Workspace.jpg>

It consists of, and can access a representation of its design as a virtual
semantic graph, which interlinks and cross-references:

   - a syntactical abstraction of the project’s source code syntax
   - documentation
   - relevant Wikipedia material
   - up-to-date metadata about the project’s development and developers

<http://3.bp.blogspot.com/_mWGkAr1Xo5Y/TVCLxXVkaZI/AAAAAAAAAzc/q1ufCKK_W7o/s1600/lod-datasets_2009-02-27.png>

It utilizes and analyzes a software library and its dependencies, to an
arbitrary stack depth, including:

   - all available “AI” algorithms, including but not limited to:
      - streaming graph data structures and graph algorithms, including
      interfaces that expose peripheral data sources and sensors as graph data
      - evolutionary search
      - high-dimensional and spatiotemporal learning and belief networks
      - probabilistic logic and inference
      - graph and N-dimensional pathfinding
      - dimensional embedding and graph drawing
      - natural language parsing and generation
   - user-interface systems
   - and other components and utilities

<http://2.bp.blogspot.com/_mWGkAr1Xo5Y/TVCL4MhN4rI/AAAAAAAAAzg/xhz_0CluCiE/s1600/decshot.png>

A networked 2D/3D zooming user-interface with 3D physics simulation features
interactively represents arbitrary information detail in seamless virtual
environments containing embodied agents that operate independently of
humans. It can embody itself within representations of its own structure,
allowing it to fluidly enter its own imagined reality models.

The resulting “operating system” simultaneously serves as a web browser,
communication tool, software development tool, simulation modeler, videogame
engine, and controller for arbitrary real-world machines and devices --
within a spectrum between fully automatic, semi-automatic, or manual modes.

<http://2.bp.blogspot.com/_mWGkAr1Xo5Y/TVCMiJ2denI/AAAAAAAAAzk/_HAgifgcBXI/s1600/fs.jpg>

Distribute the software and its source-code with an open-source license
allowing anyone to freely experiment with it and share results with the
system. Since the system is fully self reflective, any improvements that an
individual contributes can potentially improve everyone else’s instances of
it. Each instance of each system component will have multiple configuration
options and parameter, maximizing design diversity across the planetary
software eco-system. The system implements a global feedback and education
system, allowing users to suggest improvements, report problems, ask
questions, and receive answers about the software.

<http://3.bp.blogspot.com/_mWGkAr1Xo5Y/TVCMyx3BrVI/AAAAAAAAAzo/eLF7upkoN3w/s1600/ballLookSmall.png>

Developing intelligent cognitive agents will primarily involve 3D physics
simulation of a virtual body operating in a virtual world. Simulated agents
have precise control of robotic body parts and sensors that perceive, among
other signals, streams of virtual optical and acoustic phenomena. This
constitutes a robot design system that simultaneously evolves both hardware
and software aspects of virtual and real robots (resulting in potentially
unusual designs, for example: embedding cameras in robot finger-tips) -
including instructions for manufacturing them.  Also remember that
nanomachines, vehicles, and buildings, can also be thought of as robots in
this same way.  This is also capable of developing virtual machinery that
operates as intelligent “avatars” or “cursors” inside the virtual reality
user-interface, actively operating in cooperation with human-computer
interaction.

*seeAlso:*

   - http://wiki.syncleus.com/index.php/dANN
   - http://critterding.sourceforge.net
   - http://wiki.opencog.org
   - http://www.opengl.org
   - http://bulletphysics.org
   - http://sites.google.com/site/narswang/home/nars-introduction
   - http://neuroph.sourceforge.net/

-- 
P2P Foundation: http://p2pfoundation.net  - http://blog.p2pfoundation.net

Connect: http://p2pfoundation.ning.com; Discuss:
http://lists.ourproject.org/cgi-bin/mailman/listinfo/p2p-foundation

Updates: http://del.icio.us/mbauwens; http://friendfeed.com/mbauwens;
http://twitter.com/mbauwens; http://www.facebook.com/mbauwens

Commons Strategies Group, http://www.commonsstrategies.org/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.ourproject.org/pipermail/p2p-foundation/attachments/20110208/a2580500/attachment.htm 