About site: Artificial Intelligence/Neural Networks/Software - Neuropilot Project
Return to Computers also Computers
  About site: http://freespace.virgin.net/michael.fairbank/neuropilot/

Title: Artificial Intelligence/Neural Networks/Software - Neuropilot Project Showcases a java applet demo of a trained neural network piloting a lunar-lander type spacecraft over landscapes of various complexity.
Centibots__The_100_Robots_Project DARPA funded project using many small robots for mapping and search missions. Site includes videos and publications describing the collaborative communications architecture used by the Centibots.

Change_Sciences Provides research, workshops, and consulting for improving the user experience of web sites and interactive applications. Located in New York, USA.

RFC2518 This document specifies the Distributed Authoring Protocol, a set of headers and methods which extend the Hypertext Transfer Protocol to provide capabilities for overwrite prevention (locking), proper

3DEM Interactive OpenGL 3D viewer with flyby and animation capabilities. Reads most familiar US DEM formats and can export to common image and terrain formats. [Win95/98/ME/2000]

Training_City Applied training and consulting services in .NET, Visual Basic .NET, ADO.NET, and ASP.NET.

FreeFP_Themes Link directory of website themes, newsletter, and a section for help with installing them.

  Alexa statistic for http://freespace.virgin.net/michael.fairbank/neuropilot/





Get your Google PageRank






Please visit: http://freespace.virgin.net/michael.fairbank/neuropilot/
  Related sites for http://freespace.virgin.net/michael.fairbank/neuropilot/
    Pozicom_Technologies,_Inc_ EDI Pro EDI package supports ANSI X-12 4010 and 4020 standards.
    Mega_Hosters Offers shared hosting, virtual and dedicated servers, co-location and domain registration.
    Coyote_File_Transfer_System Coyote is a powerful replication system that offers secure data transfer that traverses specified ports in your firewall, and does not expose your sites to the known dangers of using NetBios and FTP.
    Configuration_Management_Kit A kit of software and resources for the mapping and management of business and IT resources
    Modelling_Autonomous_Agents_in_a_Multi_Agent_World_(MAAMAW) 2001, May 2-4, Annecy, France. European workshop on multi-agent systems. (May 2, 2001)
    Advanced_Digital_Media,_Inc_ Media conversions, streaming media technologies, branded players for content distribution, web site design and video compression.
    NoSound_net Free unlimited webspace, unlimited email, subdomain, 3 gigs for MP3 storage/share.
    Talking_with_Terry_Winograd Interview. ACM: Ubiquity. (July 23, 2002)
    perljvm The Perl to Java Virtual Machine (JVM) compiler. [Open source, GPL]
    PearlDoc Modular, web-based solutions able to cover the lifespan and flow of docu-centric information within a company; capable of managing structured and unstructured data - DM, versioning, imaging, ERM/COLD,
    Exodus Homepage of DC PAK of the Exodus demo group.
    TIOBE_Programming_Community_Index__TPCI Updated monthly. Shows language popularity via world-wide availability of skilled engineers, courses, 3rd party vendors. Calculated from search engines Google, MSN, Yahoo.
    RFC_0265 The File Transfer Protocol. A. Bhushan, B. Braden, W. Crowther, et al. November 1971.
    RFC_2868 RADIUS Attributes for Tunnel Protocol Support. G. Zorn, D. Leifer, A. Rubens, et al. June 2000.
    I_Want_Newsletters A large collection of free newsletters sorted by category.
    LightsOutSports_com Official sports screen savers, including college conferences and teams, pro and fantasy football.
    Pig_and_Cow_Design San Diego-based web design and development; includes print and software design.
    Filnet,_Inc_ Offers web development and content management services.
    Billock,_Brent Designer specializing in animated sites and components using Macromedia Flash, located in Chicago, Illinois, United States.
    Web_Traffic_Help Purchase web site traffic, including targeted traffic.
This is websites2007.org cache of m/ as retrieved on 2008.08.30 websites2007.org's cache is the snapshot that we took of the page as we crawled the web. The page may have changed since that time.
Neuro-pilot Demo

NEUROPILOT PROJECT

Copyright Mike Fairbank 2001-2008 Introduction This is a project to try to make an artificial neural network learn something challenging and interesting.  I have been using a series of “Lunar Lander” type scenarios of progressive difficulty as a motivation to develop advanced learning algorithms. Problems and Demos All of the demos below consist of a spacecraft that aims to fly to a landing pad. The spacecraft is controlled by an artificial neural network. The flying strategies you see in the demos have not been explicitly programmed by a human programmer, but have been learned by the computer autonomously, i.e. this is an example of artificial intelligence or reinforcement learning. For example, before any learning took place, the spacecraft would just fall straight to the ground and crash badly; but after crashing many times, eventually the neural network adapted itself to adopt some strategy capable of landing properly. As you can see, the strategies evolved in some cases are quite complex and would be fairly challenging for a human programmer to have invented or developed. Neuropilot1:  A 2D Problem in which a spacecraft starts in random position, orientation and motion, and has to land gently on a fixed landing pad in fuel-efficient manner.  Uses a value function approximated by a feed-forward neural net. Neuropilot2:  A 2D Problem in which a spacecraft starts from fixed position, orientation and motion, and has to find, using a range finding scanner, the landing pad and fly to it and land gently in fuel-efficient manner.  Uses a recurrent neural net.  Does not use value function. Neuropilot3:  The spacecraft tries to fly to the bottom of a tunnel using two scanners for guidance.  Uses a recurrent neural net.  Does not use value function. All demos contain a pre-trained neural network.  No learning takes place in the demos.  The sequence of flights/landscapes is pseudo-random, producing the same sequence each time.  The first few in each demo were considered during the training process, so performance is noticeably slightly better on these first cases. Background information These lunar-lander problems are from a series of problems I have been using that show progressive difficulty to solve.  These started with a very simple one-dimensional version of the lunar lander, and work up to the most complex 3rd demo. The simple one-dimensional version is demonstrated, with source code and learning algorithm, in the web-pages explaining my paper. I believe when doing a project like this you need to have a number of intermediate milestone projects that are related and form almost a continuum of difficulty. That way when one project goes wrong you have some chance of figuring out why. I think there is not much point diving straight in and attempting your final project immediately. The idea was to work through these milestone problems (and then hopefully further) until I found one that posed a challenge to the theory.  I was also hoping that I would have, by then end of this sequence of problems, a system that could be classified as in some way as “intelligent”.  I am not sure how close I have come to having achieved this. I have a full discussion of this issue here. Any comments would be welcome.  I think if I had the final demo working on 100% of all landscapes then that would be quite good. Neural Networks - Supervised Learning Generally, the neural networks of the type used here take some information in, process it, and pass some information out.  For example, in an optical character recognition system a network may take the ‘1’s and ‘0’s of a black and white pixellated image as its input, process it, and produce as its output a single number from 1 to 26 to say which letter of the alphabet it has identified.  The ‘intelligence’ behind the neural net is in the fact that you can teach it just by giving it examples (e.g. of the form, “here are 100 different letter ‘A’s, here are 100 different letter ‘B’s, etc”), and that once it has been trained, it can make reasonable guesses on input it has never seen before.  This is usually considered difficult for computers to do but straightforward for a human being. Reinforcement Learning This project is an example of a reinforcement learning problem.  Here we give a rule on what we want the net to achieve (i.e. we want it to land gently over the landing pad in a fuel-efficient manner) without specifying how.  During the training of the net, we run many test flights (which initially tend to crash badly), and for each case a ‘reward function’ is calculated based on the above criteria.  At each iteration of the training process we adjust the connections in the neural net so as to try to maximise the reward function, and hopefully after a while the net will learn to fly the spacecraft intelligently.  Reinforcement learning problems are generally harder than those that work by showing examples (supervised learning).  The reinforcement learning is never shown the correct answer - it has to work it out for itself.  This creates a significantly non-trivial challenge, because the net takes many decisions during the course of the flight but only at the end does it find out how well it did overall.  The difficulty is in working out which individual actions were responsible for doing well, or conversely badly, at the end. Learning Algorithms I have had to develop some advanced learning algorithms to produce these demos.  All learning took place off line, and no learning takes place in the demos. Reinforcement Learning by Value Gradients is the theory I developed to create the first demo. The second and third demos were created using a recurrent neural network learning algorithm that I am hoping to write up soon. Java Installation: These demos require Java to run, so if they don't work then either... Check you have Java enabled. Or install the java plug-in from Sun. Go to www.java.com and then click on the yellow "get it now" button. Contact Me If you have any feedback / questions on this then please e-mail me (see home page). To leave comments, you can also write on the Forum. Back to home page A.I. Links Here are some excellent AI theory and neural-network resources. Reinforcement Learning Repository at UMass, Amherst Backpropagator's Review Bookmark with: Delicious Digg reddit Facebook StumbleUpon What are these?  var gaJsHost = (("https:" == document.location.protocol) ? "https://ssl." : "http://www.");document.write(unescape("%3Cscript src="http://freespace.virgin.net/michael.fairbank/neuropilot//" + gaJsHost + "google-analytics.com/ga.js' type='text/javascript'%3E%3C/script%3E"));var pageTracker = _gat._getTracker("UA-4137509-1");pageTracker._initData();pageTracker._trackPageview();
 

Showcases

a

java

applet

demo

of

a

trained

neural

network

piloting

a

lunar-lander

type

spacecraft

over

landscapes

of

various

complexity.

http://freespace.virgin.net/michael.fairbank/neuropilot/

Neuropilot Project 2008 August

dvd rental

dvd


Showcases a java applet demo of a trained neural network piloting a lunar-lander type spacecraft over landscapes of various complexity.

Rules

© 2008 Internet Explorer 5+ or Netscape 6+
Recommended Sites: 1. Arts - Business - Computers - Games - Health - Home - Kids and Teens - News - Recreation - Reference - Regional - Science - Shopping - Society - Sports - World Miss Gallery - Top Anime Hentai - DVD rental by mail - Free eBooks Download - Premade MySpace Layouts - Internet Advertising - Internet Advertising - Remortgages
2008-08-30 03:42:31

Copyright 2005, 2006 by Webmaster
Websites is cool :) 126Jastrzbia Gra - Hotel Hamburg - Reklama - Hotel Brussels - Implanty