NanoBots
Stop the OctoVirus with NanoBots!
..........................
The year is 2012 and life on earth is about to end...
Back in 2003 an elite group of scientists working in a government lab at Los Alamos discovered the 'fountain of youth' for mankind. After years of very promising results they had finally succeeded in creating a DNA modified virus that was capable of rejuvenating damaged human cells. For three years OctoVirus was tested on thousands of terminally ill subjects with one hundred percent success. No disease was a match for the healing powers of this new creation! During the following seven years OctoVirus doses were administered to eighty five percent of the earth's population. Once in the bloodstream, OctoVirus maintained the human body in ‘top notch’ condition and was passed on to new generations. Scientists predicted life spans of near seven thousand years and with our new found longevity mankind began preparations to populate the galaxy.
Then things went to hell...
OctoVirus has mutated and people are dying. The lab had a couple years advanced notice that something was going terribly wrong but realized that only panic would result from notifying the masses without a way to stop the mutation. OctoVirus had been designed to require a very small number of red blood cells from its host in exchange for the near immortality it gave, this was considered a ‘fair trade’ by most of humanity. In 2010 the lab began to notice a very slight change in the coloration of red blood cells, OctoVirus was ‘improving’ its food supply by attaching a strand of mutant DNA! There was only one way to attempt to stop this disaster.
NanoBots...
In 2002, before Palm OS and the QMAP Molecular Processor made all other computers obsolete, a small lab in Arizona was developing the M-5 Neural Net Computer. M-5 provided the lab a means to train and download simulated neural network intelligence to a new breed of robotics nano-scale machines. NanoBots had been conceived as a species of injectable robots that could repair damaged cells but with the creation of OctoVirus most development on NanoBots and Simulated Neural Networks had come to a halt.
November, 2010...
Phone rings at TCL... "Gary, this is Doctor Cantor at Los Alamos, we've got a problem down here!"
That was two years ago. TCL and Los Alamos have been working on the problem since that day. Realizing that twenty years of programming would be required to bring Neural Network technology back up to speed, a plan was devised to enlist the processing power of the entire Palm community. To that end, we have put together a simulation of the problem to run on the worlds newest Super Computers (Palm OS QMAP machines, backward compatible with all Palm Color OS machines). Now the only hope for humanity is to teach the tiny NanoBot to mount a killing assault on OctoVirus and scrub the foreign DNA from damaged blood cells.
The NanoBot training simulation is being made available through the worlds largest handheld software distributor in order to reach the planets most creative minds before it's too late. There is a lot more to this story but that will have to wait, time's running out...
M-5 has eight inputs, a hidden layer and an output. Here's how we mapped the Bot simulation for training...
A NanoBot can see the eight locations surrounding it, we assign these locations to the eight M-5 inputs.
1=upper left. 2=above. 3=upper right. 4=left. 5=right. 6 =lower left. 7=below. 8=lower right.
Things it can see are represented by values.
1=edge of simulation area. 3=another bot. 5=empty space. 7=Virus. 9=blood cell.
Things it can do.
1=move up. 3=move left. 5=not move. 7=move right. 9=move down.
Things that happen.
If it moves to a Virus it eliminates the Virus and manufactures another NanoBot.
If it moves to another NanoBot it eliminates the NanoBot it moved to.
If it moves to a blood cell it repairs the cell if damaged.
Training.
The registered version has the option to automatically record training facts and also includes the registered version of M-5 Neural Net Computer free.
For the unregistered version do the following...
To create the M-5 NNtrain file inputs draw grids (like tic tack toe) on a sheet of paper. Locate your NanoBot in the center square and draw a situation in the other eight squares.
Make the Training Fact from the numbers representing the items in the 'situation'.
Determine what the NanoBot should do in this situation and use that number as the Desired Output for the training fact.
Example:.
empty empty cell
empty [Bot] cell
empty empty empty
The Training Fact would be 55959555 7 ( in this situation, move to the right )..
See the M-5 documentation for NNtrain memo details.
After you have created many training facts, use M-5 to train the NanoBots.
Use the M-5 'Run' mode to see how well your Bots handle the OctoVirus.
Tip... Copy the NNtrain file and paste it into a new file. Change the new file's top line to NNrun. You can 'Run' M-5 with this file to get an NNout file showing how well M-5 is learning each fact. If there are some facts that M-5 is having a hard time learning it may be that they conflict with other facts you are teaching, try changing the fact slightly and training some more (remember to update your NNrun file).
The NNweights file produced by M-5 is the Bots 'brain'. Change the top line of the file to NNbot.
(Be sure to delete or change the name of earlier versions of NNbot).
NanoBots will respond as trained 75% of the time and improvise 25%.
OctoVirus evolves by deriving 25% of its intelligence from the NanoBots actions.
Version 0.9
Install NanoBots.prc and mathlib.prc (and M-5.prc if you want to train them).
Copy the NNbot file to the MemoPad .
Launch NanoBots.
Tap RUN.
Answer the questions.
Tap a blood cell.
Tap RUN.
Hold the Up or Down button to stop the program, even if it appears that nothing is happening you still need to hold the button to stop.
To run the program again, exit and re-launch NanoBots.
The first person to email TCL a NNbot memo that consistently eliminates OctoVirus and repairs damaged cells will receive a full refund of their registration fee and may go down in history for saving the world!
Visit http://www.talestuff.com/ for all your Palm accessories including Emergency Chargers and Ir Blasters.
With a little imagination there are a variety of 'real world' applications for Neural Nets!
United States Patent 4,970,819 Mayhak. November 20, 1990 ---------------------------------------------------------------
Firearm safety system and method.
Abstract: Actuation of the firing mechanism of a firearm is prevented until grip pattern sensing means on the handgrip of the firearm supply to a microprocessor signals corresponding to a grip pattern stored in a programmed simulated neural network memory. All of these components are contained within the firearm. Programming of the neural network memory is accomplished by using a host computer with a simulated neural network to train that network to recognize a particular grip pattern using grip pattern signals generated by the grip pattern sensing means as the sensing means is repeatedly gripped for the person for whom the firearm is to be programmed. -------------------------------------------------
Inventors: Mayhak; Gary D. (Scottsdale, AZ) Assignee: V/GER, Inc. (Scottsdale, AZ) Appl. No.: 411636 Filed: September 25, 1989 Current U.S. Class: 42/70.01; D22/104 Intern'l Class: F41A 017/06 Field of Search: 42/70.01,70.11
What are Neural Nets? Look here...
http://blizzard.gis.uiuc.edu/htmldocs/Neural/neural.html
http://www.dacs.dtic.mil/techs/neural/neural3.html
http://www.enee.umd.edu/medlab/neural/nn1.html
Gary Mayhak / Tech Center Labs / http://www.talestuff.com/
last updated 12-02-02