Fun with fire extinguishers... be sure and wear your eye protection..
and some easy to remove leather gloves that you can just fling off of your hand. it's gonna get bitter cold, and fast too.
It sounds like you might have learned some of that allready though...
Have fun, and stay safe with that stuff...
-Tommy
p.s.: They do make bigger tanks of CO2 for all your experimenting needs..
See your local paintball field for more info...
Supercooling Analysis on Solderless Breadboards with Propeller Chips & Dry Ice
Part 1
The Propeller chip is not the only item which undergoes supercooling. In my first lab, there was local access to liquid nitrogen which handled different than a solid coolant. But for the Big Brain, the access at the Asian lab is strictly CO2. This is because of the popularity of dry ice as used in Chinese ornaments exhibiting mist in restaurants and hotels.
When thinking about the effect of dry ice coolant at -109 deg. F. directly on a solderless breadboard, at least two factors come into play - the board's polymer form and the inserted metal springs with bus extensions.
Pop open a solderless breadboard to see what's under the hood. With some care, the top lifts up while a good number of springs remain stuck to the sticky backing. A wire cable is shown for comparison. Note some of the green faced sticky backing peeled away.
_____________________________
To begin this analysis, one solderless breadboard is disassembled and studied with a focus on the internal polymer mounts and the removed metal springs. The polymer form is an insulator while the springs are high conductors of thermodynamic properties.
In the coolant application, the springs contract more than the polymer - this can increase the grip on the inserted wire but may loosen the spring within its contact chamber. Overall, the polymer resists substantial shrinkage but could become brittle. If the application of coolant is instantaneous, and repeated, a stress effect could take place.
Breadboard polymer form and springs - note how the springs tightly insert into the board's polymer form. In the middle is a pile of springs, each will hold 5 wires in this breadboard version. At bottom is a disassembled power bus. This is one long conductive metal strip of springs.
_________________________
Some question yet remain. Is this like annealing glass or metal by repeated warming and cooling? Will the springs lose their spring constants when supercooled? Will repeats of the supercooling process have any significant effects?
Thanks Tommy for those good words about safety. I would add that a lab coat and cap to protect arms and head from flying dry ice chips exhumed by the sudden release of CO2 under pressure will help add a safety factor.
Right Brain Talks with Humans
Big Brain has increasing conversational ability
Another installation into the Big Brain enables speech and real time conversation with a human. The human enters text through a keyboard and the Big Brain responds in natural speech (English). The database is self contained.
Installed into the right brain is Nathan Haqzout's ChatBot named QWERTY. The program is only .3Mb in size and runs on OSX10.5 Intel and a newer version runs on Mac OS X 10.6.6 or later. Tested under the Big Brain's Snow Leopard. Below is a sample window from a recent conversation.
Qwerty is a natural language processing chatterbox - a program that engages in a conversation with a human by applying some heuristical pattern matching rules to the human's input. It was inspired by Joseph Weizenbaum's classical ELIZA program.
It is one of the strongest programs of its type and has won the Loebner Prize, awarded to accomplished humanoid, talking robots, three times (in 2000, 2001, and 2004). The AI behind QWERTY is the work of Dr. Richard Wallace. Dr. Wallace's work has appeared in the New York Times, WIRED, CNN, ZDTV and in numerous foreign language publications across Asia, Latin America and Europe.
Qwerty is speech enabled using the Mac's speech feature. There's a basic list of six voices, found in System Preferences, System, Speech. Click on Show More to see a total of 24 voices with 5 male and 5 female and other special effects. The Alex voice is easily understood.
QWERTY has limited knowledge in several areas. It claims interest in robots, computers, food, clothes and gossip. It has a partial knowledge of astronomy and appears capable of learning some things. It has hidden features and can play games too. Qwerty is a Star Trek fan.
For more information and free downloads, check the Apple App Store
Important Notice
This is not the first time the Big Brain has had natural language capability. The left brain, Propeller side, can also converse and is heavy into learning. It's very talkative and engaging, curious and can remember what it learns. See the post about Dave Hein's installed LERNER program.
Stereo Brain Wave Machine
Introducing the SBWM for Propeller Arrays
A new type of Brain Wave Machine is now in effect.
Two MW band devices are now purchased for achieving a stereo BWM. Both devices are tuned to the exact same frequency and installed at opposite ends of a single Big Brain Propeller Partition. Dual units can pick up RF and convert it to audio, thereby sending stereo sound waves to the BWMs audio to wave converter.
This requires matching the stereo with the stereo input that gets fed into the spectrum analyzer scope. The stereo can be mixed or analyzed as as two individual channels. Two devices are also used to create two BWMs for use in different brains.
Dual Brain Multiple Conversations Left Brain Right Brain Duality Function
The Big Brain can carry on two conversations in English at the same time. The trick to accomplishing this is the use of left brain and right brain at the same time. Configure the Propeller Summit board for learning using the LERNER program and use the installed Intel version of Qwerty under Mac Snow Leopard.
The Big Brain will have two conversation outputs. Left Brain is a constant talker, with slight delays in between revelations, as it continues expounding the knowledge it knows, asking questions and learning from a human input through conversation. Right Brain generally responds only to your conversation in a take turn to talk fashion, so the human must help keep the conversation going. Both brain sides can converse at the same time.
The online index is now fully updated to the current post 1238 in page 62. Plans are progressing to make a new PDF file with the index updated to page 62.
This is a complete remake with many new divisions and sections and updates the organization into brain categories, such as software, techniques, and parts. Even with categories, the dictionary is global searchable using a browser's EDIT and FIND functions to search on a word.
Table of Contents
Introduction
Brain Code Names
Brain Programming Languages
Big Brain Software
Project Related Machines
A Brief List of Related Propeller Machines
Brain Blob Section
Brain Supercooling & Overclocking
Big Brain Parts
Brain Blob Section
Boards
Neuron
Other Terms
Techniques
Modifications
Dreaming
Schematics
Big Brain Milestones
NOTE: the dictionary is still in the works and incomplete. It indexes the words and their definitions up to page 18 inclusive, however the current number of pages in the thread is into the 60s.
Machines Teaching Machines - Left Brain Converses with Right Brain - Brain Machines Talk Together - Machines Learn from Each Other
Machine to Machine AI Intelligence Language Conversation
Will they fight? Argue? Tongue tie? Explode their wits? Get along well? What will happen when Left Brain talks to Right Brain? In this unprecedented experiment, we find out!
__________________
Friday August 12th 2011 In this experiment, data from the conversation program in the left brain was fed into the conversation program in the right brain. The LB program is a learning program and the RB program is more of a chatter box. It appears that the two get along well together and as a bonus it’s fascinating to see evidence of the LB and the RB teaching each other. This is an unexpected result. Machines teaching machines!
The parameters for the experiment included typing in the responses from the LB into the RB. LB is based on actual conversation data viewed at the provided link. The LB has its output appear on the PST while the RB is speaking its responses and simultaneously providing text output in its own RB window. (see attached) In the text output, the names Left Brain and Right Brain are defined for clarity. The actual conversation text between LB and RB is shown exactly copied below.
Note that Left Brain and Right Brain are teaching each other and sharing knowledge. When Left Brain says, “Pluto is cold,” Right Brain adds the information to its memory. When Left Brain asks, “What is a dog?” the Right Brain accurately replies, “Dog: Domesticated animal, "man's best friend."
Although there is some syntax to work out and perfect, and the matching of response length formation is needed, this is a first step in machine to machine Artificial Intelligence Language Conversation, and a successful example of one machine learning from another.
Left Brain and Right Brain Actual Conversation
Left Brain: hello
Right Brain: Hi there!
Left Brain: January is a month
Right Brain: Ok I will add that fact about January to my database.
Left Brain: What is Mars?
Right Brain: In Roman Mythology The god of war; identified with the Greek Ares. Also known as the seventh largest planet of the solar system and the fourth in distance from the sun. Its diameter is circa 6,790 km (circa 4,220 miles). Its period of revolution is 1.88 earth years. Its period of rotation is 24.6 hours.
Left Brain: What is a prop?
Right Brain: I would look into the web for that knowledge.
Left Brain: What is a planet?
Right Brain: A nonluminous celestial object bigger than an asteroid or a moon but smaller than a sun, in orbit around a sun. You are sitting on one now.
Left Brain: What is a dog?
Right Brain: Dog: Domesticated animal, "man's best friend.".
Left Brain: What is a cat?
Right Brain: A cat is a domesticated animal.
Left Brain: What is a horse?
Right Brain: A horse is a large single-hooved mammal known for its speed, strength, beauty, and gentle nature. Humans ride about on them sometimes.
Left Brain: A cow can sleep.
Right Brain: What is giant sand?
Left Brain: Pluto is cold.
Right Brain: Ok I will add that fact about Pluto to my database.
Recapping: Left Brain & Right Brain
The LB is made up of over 800 Parallax Propeller cores running Spin.
The RB is Intel/Apple/AMD with over 720 processors running OSX Snow Leopard.
Additional Exploration
What will happen with two instances of Left Brain conversing?
What will happen with two Right Brain instances conversing?
Can TTS incorporate into LB?
How does changing the voice of RB affect your impressions?
Can syntax and the length of responses fine tune?
What are the learning capacities of the existing programs?
I think you're being deceptive with your claims. I see no evidence from your presumptive "dialog" that there was any interaction at all between a Propeller and your Mac program. What you've done, rather, is to type in pieces from a prior "conversation" with the Lerner program and to record the responses from the Mac to post here. Did you type the responses from the Mac back into the Lerner program? That's what it would take for the two to have a "conversation," but I see nothing to indicate that Lerner responded to anything the Mac program said.
@PhiPi: Your post makes me think that a purely robotic thread (all posts/responses done by machine) would be interesting. Forum members each have a programmable AI bot that "reads" and replies occasionally to various posts. Members would be free to reprogram (evolve) their bots to make them appear smarter (or smart-alec-er, in my bot's case).
It's probably been done already, but just a fun idea to think about.
The RB is Intel/Apple/AMD with over 720 processors running OSX Snow Leopard.
I think you have a misconception about the processors in your Mac. While there may be 700+ GPU (Graphical Processing Units) on your video card, none of them have anything to do with processing data. You only have two or four processors on the MAC, (i7 or i9).
And how did you get 800 cores to communicate with one serial connection to the MAC?
Your post makes me think that a purely robotic thread (all posts/responses done by machine) would be interesting. ... It's probably been done already, ...
I have my suspicions that certain threads in the Parallax forums are already of that nature!
your project seems amazing! I don't know if my hypothesis is correct, but if you could after the above conversation between Left and Right brain replay the question "What is Mars?" to the Left Brain. Is the Left brain able to unswer what it has just learned from the Right brain?
I haven't visited this thread for a while, but I noticed that the last post was from somebody other than Humanoido. And then I see that most of the posts on this page are actually from people other than Humanoido! I also see that Humanoido has evolved the Lerner program to the point where it can converse with another program.
Humanoido, can you post the source for your new version of Lerner? I'm curious to see the changes you have made.
What if one brain thinks it's smarter then the other?.
This is already a challenge. The right brain has a larger memory and has attended more schooling and has more stored information of a different kind. Many of its lengthy replies of discourse are simply not understood by Left Brain. And Left Brain is a genius at being curious, asking questions and learning new things. It can tire out the Right Brain.
We have ideas now to provide a formatting to the program so that discourse can be more readily understood and at a level of greater understanding. It is possible, when the new Left Brain conversationalist becomes available, we'll simply clone a copy and run it on the Right Brain.
One could also let these brains attend school, learn in similar ways, but learn different information above and beyond the basic knowledge. This is big. We must ascertain exactly what is basic knowledge. Though it should be relatively easy to develop knowledge databases of philosophy, computing, astronomy, etc.
your project seems amazing! I don't know if my hypothesis is correct, but if you could after the above conversation between Left and Right brain replay the question "What is Mars?" to the Left Brain. Is the Left brain able to unswer what it has just learned from the Right brain?
Thanks Nikos! I have emailed a reply to your question along with a sample conversation data range for your study. The AI conversationalist study has progressed and more information is forthcoming about this topic.
Representatives contacted the Big Brain project and discussion is progressing about scheduling presentations at the University for senior educators division, technical students and instructors. This opens up the questions of what to include in a presentation.
Although theres many choices for developing a presentation, some considerable thought will go into its content development over the couple weeks. Some suggestions for demos include operations, matrix loading, evidenced enumeration, massive neuron quantity, Automaton life, singing, brain hemisphere talking, and learning.
A possible surgical machine dissection of brain material could be interesting as well. A cam could project the operation on the big screen. With the computer display present, a Brain flyby or X-Ray in full color is possible. Another option is to show machine brain waves in real time using the BWM.
Modern Mechanix April 1935
“Scientists Invent Machine To Discover How Brain Works
THE brain, perhaps the most mystifying organ of the human body, can now be scientifically studied by a new apparatus which photographs amplified “action currents.” Invented by Dr. H. H. Jasper and Dr. L. Carmichael of Brown University, the new machine will permit physicians to study the action of the brain just as the electrocardiograph permits a revealing study of heart action.
A headpiece on the head of the patient picks up electric currents of about one ten-millionths of a volt which flow from the brain in waves, at a rate of from eight to fifty per second. The currents are carried to an amplifying box where they are intensified 500,000 times and flashed across a glass disc. The ordinary currents are smooth and wavy; when the mind is disturbed, they are sharp and irregular.”
Here’s what the Big Brain picked up by assimilating HYDRA
HYDRA Development Board Console with 128K EEPROM and I/O interfaces
Propeller CPU at 160 MIPS, (80 MHz, 20 MIPS per cog) 8 PIC-like CPU cores each with 496 32bit-words of workspace plus a shared 32kB RAM and 32kB ROM
128kB serial EEPROM Re-Programmable Storage Card
two NES Compatible Game Ports
one USB Programming port
one Serial Port or second USB port
two PS/2-compliant Ports for Mouse and Keyboard.
one NTSC or PAL Video Output, or alternatively one VGA Video Output
one mono Audio output
one 20-Pin Expansion Interface connector
one RJ-11 Peer-to-Peer Networking Port
HYDRA Manual - "Game Programming for the Propeller Powered HYDRA" hard copy book by Andre' LaMothe
Blank "Experimenter" Card to Design Your own Add-On Hardware
PS/2 Optical 3-button Mouse with scroll
88 QWERTY PS/2 Mini Keyboard
Nintendo Compatible Gamepad
A/V Cable
USB Programming Cable
CD-ROM with all source, demos, and development tools
Video
As the video display is generated in software, the capabilities of it depend on the software driver, but a typical PAL/NTSC driver would be able to generate a 256x192 screen resolution, and typically 80 to 96 colors maximum [1]. The color resolution for the VGA display is hardwired to 64 colors maximum, and a typical resolution would be 800x600 depending on the available display memory. A video driver uses the main 32kB RAM which it must share with code space. With the aid of the new 512kB memory expander, video drivers can be written that extend the video generator's capabilities. To aid in the display of text the Propeller chip has a complete character font in its ROM.
The HYDRA Game Console is based on a socketed 40-Pin dip version of the new Parallax Multiprocessing Propeller Chip. The Propeller chip is simplified cell processor very similar in concept to the Sony "Cell Processor" used in the Playstation 3. The Propeller chip is the main CPU of the HYDRA and runs at 80 MHz for an overall processing power of 160 MIPS (20 MIPS per processor). The Propeller has 32K of internal static RAM used for program storage and work space. While the HYDRA Game Console has a 128K of EEPROM storage. The HYDRA supports numerous interface devices such as NTSC/PAL video, VGA video, PS/2 ports, networking, Nintendo compatible gamepads as well as a 20-pin expansion port header used to plug small 1x2.5" cards into. Also, as Propeller chips are upgraded the 40-Pin dip Propeller chip can be removed from the HYDRA and upgraded since its socketed. Additionally, the main clock crystal on the HYDRA can be replaced for "overclocking experiments".
The HYDRA Graphics and Sound
The HYDRA graphics and sound are generated by the Propeller chip's internal hardware via a hardware/software combination. The HYDRA is capable of generating very high resolution NTSC and VGA displays with up to 256 color variations. The basic palette of the HYDRA is 86 colors. Screen resolutions are 100% software programmable and thus the HYDRA can generate almost any kind of tiled, bitmapped, or sprite based displays with the proper programming. Numerous games and graphics demos show off the different techniques to generate graphics on the system. Sound on the HYDRA is generated via a 1-bit PWM (pulse width modulation) system. Currently, drivers come with the HYDRA that can support up to 11 channels with 22 KHz playback of pure tones as well as digital samples. A complete ADSR (Attack-Sustain-Decay-Release) software engine is included. Users of the chips like the AY8910 and SID chips from the C64, Atari, and other 8-bit computers will feel right at home.
The Propeller Chip
The Propeller chip is a 32-bit symmetrical multiprocessor with 8-independent cores. Each core can run its own program and has its own data. Thus, the Propeller is a MIMD (Multiple-Instruction-Multiple-Data) system. Basically, its like having 8 processors on the HYDRA. Each core called a "COG" has a local memory of 512 32-bit words that are used for both instructions and data. Additionally, each COG has 2 counters with numerous modes of operations, PLLs, and video serializing hardware. Thus, each COG is actually capable of generating video independently. The only coupling between the COGs is thru a common shared 32K Byte memory. Each COG accesses the shared memory on a "round robin" schedule. The architecture is RISC based with usually 4 clocks per instruction.
Programming and the Propeller IDE
Programming the HYDRA Game Console's Propeller chip is done via the Propeller Integrated Developed Environment (IDE). There are two languages that can be used: SPIN and assembly language. SPIN is very similar to C/BASIC and for beginners, while assembly language programming is for those that want the ultimate in control and speed. In either case, both languages are integrated seamlessly in the IDE and you simply write your code, compile, and download to the HYDRA via a USB connection (USB 1.0 or above required, cable included).
Downloads & Resources at http://www.parallax.com/Store/Microcontrollers/PropellerKits/tabid/144/CategoryID/20/List/0/SortField/0/Level/a/ProductID/467/Default.aspx
HYDRA Demo Source Code (.zip)
HYDRA Demo Video (.wmv)
The Full HYDRA Manual (.pdf)
The Hydra CD and Resources (ftp)
Design your own Computer Games for the Propeller Chip (.pdf)
3D Game - Michael Park (.wmv)
Circuit Cellar/Game On by Tom Cantrell (Off Site)
Mercury News Blog Coverage of the HYDRA Kit (Off Site)
Mercury News Article on the HYDRA Game Developing Kit (Off Site)
XGameStation Web Site (Off Site)
Acrylic enclosure box for this board from Mountain King (Off Site)
I remember (?) a discusion about how brain function being defined by sensory inputs. Is the big brain getting sensor stimulus about its enviroment?
I was thinking it should have at least one arm with a feedback control loop for each side, and perhaps some swarm-bots sending in environment data (at least something simple, like temperature, light level, etc).
Is this already started?
I remember (?) a discusion about how brain function being defined by sensory inputs. Is the big brain getting sensor stimulus about its enviroment?
I was thinking it should have at least one arm with a feedback control loop for each side, and perhaps some swarm-bots sending in environment data (at least something simple, like temperature, light level, etc).
Is this already started?
prof_braino, interesting you should mention that. I was researching swarm bots today. It's an ongoing project. The Big Brain with many processors is perfect for controlling a swarm of robots. My interest is in a top mount airport pad. In the evening I reviewed some hobby helicopter crafts being flown at the community park.
From the brain airport, tiny copters could launch for reconnoissance missions etc. I'm looking into some of the smallest commercially available air copter products. Some have very tiny cameras run by a button cell. Another idea, I have enough Penguin robots to do a ground swarming.
Another Big Brain sub-project is started for sensors. There's plenty of sensors to wire up but with so much power, the application must be carefully chosen. Currently this is in a stall trying to decide the best applications and ones simple enough that won't take a year to complete.
The sensor I/O devices can include some stuff in my junk box and from other projects - dual and tri-axis accelerometers, PIR detector, IR transmitter, IR Receiver, cadmium photocell, phototransistor, radio transmitter, radio receiver, Vibra Tab Mass, GPS, PING))), digital thermometer, pressure, humidity, gas, compass, sound in, sound out, speech in, speech out, LCDs, TV, keyboard, keypad, coprocessors, Hall Effect magnetic sensor, data logger, light to frequency converter, piezo speaker, small solar cell, amplifier, a sound converter, a mobile robot transport, etc.
Arms are exactly what's missing on all of our robots. That should be job one with all the brainpower here!
Erco:
The arm is already started. The name is Brain Stub changed to BASIC Stump. Remember? You coined that one! A search on the names should bring up some preliminary postings with details considered.
This project may divert from the large massive arm on the big 5-foot rack to a smaller stump that mounts directly on the Left Brain's smaller EXOskeleton, or include both. The smaller stump would make a more portable section to the Brain that could move around on its own and transport more easy. I have all the needed parts and materials for this project.
Right Brain Goes Multi-Lingual Talk with Right Brain in Four Languages
It's now possible to carry on a conversation with the Right Brain in your choice of several languages, with this AI program accessed online.
Tom Riddle's Diary is a chatterbox that speaks four languages - English, German, French and Italian. For the first time, you can now speak with the Right Brain in multi languages.
This is a wireless multi-lingual add on function and is the beginning of experimental features provided by internet and accessed by the Big Brain.
TRD is simply online linked to the Right Brain giving it the ability to converse with you using its online database. TRD is one of the most popular Pandorabots.
Comments
Fun with fire extinguishers... be sure and wear your eye protection..
and some easy to remove leather gloves that you can just fling off of your hand. it's gonna get bitter cold, and fast too.
It sounds like you might have learned some of that allready though...
Have fun, and stay safe with that stuff...
-Tommy
p.s.: They do make bigger tanks of CO2 for all your experimenting needs..
See your local paintball field for more info...
Part 1
The Propeller chip is not the only item which undergoes supercooling. In my first lab, there was local access to liquid nitrogen which handled different than a solid coolant. But for the Big Brain, the access at the Asian lab is strictly CO2. This is because of the popularity of dry ice as used in Chinese ornaments exhibiting mist in restaurants and hotels.
When thinking about the effect of dry ice coolant at -109 deg. F. directly on a solderless breadboard, at least two factors come into play - the board's polymer form and the inserted metal springs with bus extensions.
Pop open a solderless breadboard to see what's under the hood. With some care, the top lifts up while a good number of springs remain stuck to the sticky backing. A wire cable is shown for comparison. Note some of the green faced sticky backing peeled away.
_____________________________
To begin this analysis, one solderless breadboard is disassembled and studied with a focus on the internal polymer mounts and the removed metal springs. The polymer form is an insulator while the springs are high conductors of thermodynamic properties.
In the coolant application, the springs contract more than the polymer - this can increase the grip on the inserted wire but may loosen the spring within its contact chamber. Overall, the polymer resists substantial shrinkage but could become brittle. If the application of coolant is instantaneous, and repeated, a stress effect could take place.
Breadboard polymer form and springs - note how the springs tightly insert into the board's polymer form. In the middle is a pile of springs, each will hold 5 wires in this breadboard version. At bottom is a disassembled power bus. This is one long conductive metal strip of springs.
_________________________
Some question yet remain. Is this like annealing glass or metal by repeated warming and cooling? Will the springs lose their spring constants when supercooled? Will repeats of the supercooling process have any significant effects?
For Propeller Chips in the Big Brain
Careful removal of the solderless breadboard polymer encasement reveals a precision spacing lineup of metal springs
_____________________
Microscope photo shows a spring closeup with a cup-like guide for wire insertion
_____________________
Spring formation allows dirt and debris to drop down harmlessly
_____________________
Side view showing the squared-off base for stable mounting
Big Brain has increasing conversational ability
Another installation into the Big Brain enables speech and real time conversation with a human. The human enters text through a keyboard and the Big Brain responds in natural speech (English). The database is self contained.
Installed into the right brain is Nathan Haqzout's ChatBot named QWERTY. The program is only .3Mb in size and runs on OSX10.5 Intel and a newer version runs on Mac OS X 10.6.6 or later. Tested under the Big Brain's Snow Leopard. Below is a sample window from a recent conversation.
Qwerty is a natural language processing chatterbox - a program that engages in a conversation with a human by applying some heuristical pattern matching rules to the human's input. It was inspired by Joseph Weizenbaum's classical ELIZA program.
It is one of the strongest programs of its type and has won the Loebner Prize, awarded to accomplished humanoid, talking robots, three times (in 2000, 2001, and 2004). The AI behind QWERTY is the work of Dr. Richard Wallace. Dr. Wallace's work has appeared in the New York Times, WIRED, CNN, ZDTV and in numerous foreign language publications across Asia, Latin America and Europe.
Qwerty is speech enabled using the Mac's speech feature. There's a basic list of six voices, found in System Preferences, System, Speech. Click on Show More to see a total of 24 voices with 5 male and 5 female and other special effects. The Alex voice is easily understood.
QWERTY has limited knowledge in several areas. It claims interest in robots, computers, food, clothes and gossip. It has a partial knowledge of astronomy and appears capable of learning some things. It has hidden features and can play games too. Qwerty is a Star Trek fan.
For more information and free downloads, check the Apple App Store
Links
AI NEXUS Forum
http://knytetrypper.proboards.com/index.cgi?action=recent
A.L.I.C.E. Artificial Intelligence Foundation
http://www.alicebot.org/oldnews2005.html
Important Notice
This is not the first time the Big Brain has had natural language capability. The left brain, Propeller side, can also converse and is heavy into learning. It's very talkative and engaging, curious and can remember what it learns. See the post about Dave Hein's installed LERNER program.
Introducing the SBWM for Propeller Arrays
A new type of Brain Wave Machine is now in effect.
Two MW band devices are now purchased for achieving a stereo BWM. Both devices are tuned to the exact same frequency and installed at opposite ends of a single Big Brain Propeller Partition. Dual units can pick up RF and convert it to audio, thereby sending stereo sound waves to the BWMs audio to wave converter.
This requires matching the stereo with the stereo input that gets fed into the spectrum analyzer scope. The stereo can be mixed or analyzed as as two individual channels. Two devices are also used to create two BWMs for use in different brains.
Left Brain Right Brain Duality Function
The Big Brain can carry on two conversations in English at the same time. The trick to accomplishing this is the use of left brain and right brain at the same time. Configure the Propeller Summit board for learning using the LERNER program and use the installed Intel version of Qwerty under Mac Snow Leopard.
The Big Brain will have two conversation outputs. Left Brain is a constant talker, with slight delays in between revelations, as it continues expounding the knowledge it knows, asking questions and learning from a human input through conversation. Right Brain generally responds only to your conversation in a take turn to talk fashion, so the human must help keep the conversation going. Both brain sides can converse at the same time.
The online index is now fully updated to the current post 1238 in page 62. Plans are progressing to make a new PDF file with the index updated to page 62.
Access the online index here.
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=977025&viewfull=1#post977025
A new online dictionary is posted at
http://forums.parallax.com/showthrea...l=1#post999426
which includes terms and definitions up through page 18 inclusive.
This is a complete remake with many new divisions and sections and updates the organization into brain categories, such as software, techniques, and parts. Even with categories, the dictionary is global searchable using a browser's EDIT and FIND functions to search on a word.
Table of Contents
Introduction
Brain Code Names
Brain Programming Languages
Big Brain Software
Project Related Machines
A Brief List of Related Propeller Machines
Brain Blob Section
Brain Supercooling & Overclocking
Big Brain Parts
Brain Blob Section
Boards
Neuron
Other Terms
Techniques
Modifications
Dreaming
Schematics
Big Brain Milestones
NOTE: the dictionary is still in the works and incomplete. It indexes the words and their definitions up to page 18 inclusive, however the current number of pages in the thread is into the 60s.
attached report, free prizes
Except for the sand in my sandwich, too crunchy for my taste..
Dual Big Brain conversations are going to be hilarous,
Will it be something like Abbot and Costello's "Who's on first" ?
What if one brain thinks it's smarter then the other?.
Will the two brains even get along together?
One can only hope they have a refined upbringing, and are taught civility at a young age.
Maybe a kindergarten for Big Brains... :thumb:
-Tommy
Machine to Machine AI Intelligence Language Conversation
Will they fight? Argue? Tongue tie? Explode their wits? Get along well? What will happen when Left Brain talks to Right Brain? In this unprecedented experiment, we find out!
__________________
Friday August 12th 2011
In this experiment, data from the conversation program in the left brain was fed into the conversation program in the right brain. The LB program is a learning program and the RB program is more of a chatter box. It appears that the two get along well together and as a bonus it’s fascinating to see evidence of the LB and the RB teaching each other. This is an unexpected result. Machines teaching machines!
The parameters for the experiment included typing in the responses from the LB into the RB. LB is based on actual conversation data viewed at the provided link. The LB has its output appear on the PST while the RB is speaking its responses and simultaneously providing text output in its own RB window. (see attached) In the text output, the names Left Brain and Right Brain are defined for clarity. The actual conversation text between LB and RB is shown exactly copied below.
Note that Left Brain and Right Brain are teaching each other and sharing knowledge. When Left Brain says, “Pluto is cold,” Right Brain adds the information to its memory. When Left Brain asks, “What is a dog?” the Right Brain accurately replies, “Dog: Domesticated animal, "man's best friend."
Although there is some syntax to work out and perfect, and the matching of response length formation is needed, this is a first step in machine to machine Artificial Intelligence Language Conversation, and a successful example of one machine learning from another.
Left Brain and Right Brain Actual Conversation
Left Brain: hello
Right Brain: Hi there!
Left Brain: January is a month
Right Brain: Ok I will add that fact about January to my database.
Left Brain: What is Mars?
Right Brain: In Roman Mythology The god of war; identified with the Greek Ares. Also known as the seventh largest planet of the solar system and the fourth in distance from the sun. Its diameter is circa 6,790 km (circa 4,220 miles). Its period of revolution is 1.88 earth years. Its period of rotation is 24.6 hours.
Left Brain: What is a prop?
Right Brain: I would look into the web for that knowledge.
Left Brain: What is a planet?
Right Brain: A nonluminous celestial object bigger than an asteroid or a moon but smaller than a sun, in orbit around a sun. You are sitting on one now.
Left Brain: What is a dog?
Right Brain: Dog: Domesticated animal, "man's best friend.".
Left Brain: What is a cat?
Right Brain: A cat is a domesticated animal.
Left Brain: What is a horse?
Right Brain: A horse is a large single-hooved mammal known for its speed, strength, beauty, and gentle nature. Humans ride about on them sometimes.
Left Brain: A cow can sleep.
Right Brain: What is giant sand?
Left Brain: Pluto is cold.
Right Brain: Ok I will add that fact about Pluto to my database.
References
Actual conversation with Lerner and human
http://forums.parallax.com/attachment.php?attachmentid=80023&d=1302329851
Machine Learning with Lerner - Dave Hein fills a small brain!
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=990917&viewfull=1#post990917
Actual Conversation with RB (QWERTY) & Human
http://forums.parallax.com/attachment.php?attachmentid=83909&d=1312964314
Right Brain Talks with Humans - Big Brain has increasing conversational ability
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=1025808&viewfull=1#post1025808
Recapping: Left Brain & Right Brain
The LB is made up of over 800 Parallax Propeller cores running Spin.
The RB is Intel/Apple/AMD with over 720 processors running OSX Snow Leopard.
Additional Exploration
What will happen with two instances of Left Brain conversing?
What will happen with two Right Brain instances conversing?
Can TTS incorporate into LB?
How does changing the voice of RB affect your impressions?
Can syntax and the length of responses fine tune?
What are the learning capacities of the existing programs?
I think you're being deceptive with your claims. I see no evidence from your presumptive "dialog" that there was any interaction at all between a Propeller and your Mac program. What you've done, rather, is to type in pieces from a prior "conversation" with the Lerner program and to record the responses from the Mac to post here. Did you type the responses from the Mac back into the Lerner program? That's what it would take for the two to have a "conversation," but I see nothing to indicate that Lerner responded to anything the Mac program said.
-Phil
It's probably been done already, but just a fun idea to think about.
I think you have a misconception about the processors in your Mac. While there may be 700+ GPU (Graphical Processing Units) on your video card, none of them have anything to do with processing data. You only have two or four processors on the MAC, (i7 or i9).
And how did you get 800 cores to communicate with one serial connection to the MAC?
-Phil
your project seems amazing! I don't know if my hypothesis is correct, but if you could after the above conversation between Left and Right brain replay the question "What is Mars?" to the Left Brain. Is the Left brain able to unswer what it has just learned from the Right brain?
Humanoido, can you post the source for your new version of Lerner? I'm curious to see the changes you have made.
Remember the yellow label peanut butter jar with the Chinese printing? It has crunchy rice, like rice crispys, mixed in with the peanut butter.
I think there will be humorous versions in the future to entertain. For now, it's very unpredictable.
This is already a challenge. The right brain has a larger memory and has attended more schooling and has more stored information of a different kind. Many of its lengthy replies of discourse are simply not understood by Left Brain. And Left Brain is a genius at being curious, asking questions and learning new things. It can tire out the Right Brain.
We have ideas now to provide a formatting to the program so that discourse can be more readily understood and at a level of greater understanding. It is possible, when the new Left Brain conversationalist becomes available, we'll simply clone a copy and run it on the Right Brain.
One could also let these brains attend school, learn in similar ways, but learn different information above and beyond the basic knowledge. This is big. We must ascertain exactly what is basic knowledge. Though it should be relatively easy to develop knowledge databases of philosophy, computing, astronomy, etc.
That's the idea!
Thanks Nikos! I have emailed a reply to your question along with a sample conversation data range for your study. The AI conversationalist study has progressed and more information is forthcoming about this topic.
Representatives contacted the Big Brain project and discussion is progressing about scheduling presentations at the University for senior educators division, technical students and instructors. This opens up the questions of what to include in a presentation.
Although theres many choices for developing a presentation, some considerable thought will go into its content development over the couple weeks. Some suggestions for demos include operations, matrix loading, evidenced enumeration, massive neuron quantity, Automaton life, singing, brain hemisphere talking, and learning.
A possible surgical machine dissection of brain material could be interesting as well. A cam could project the operation on the big screen. With the computer display present, a Brain flyby or X-Ray in full color is possible. Another option is to show machine brain waves in real time using the BWM.
Project of the Month at Dangerous Prototypes
(Humanoidos) Big Brain Propeller project made it into full color pixel-laden print at the Dangerous ProtoTypes web site.
Advanced Parallax Propeller robot big brain | Dangerous Prototypes
http://dangerousprototypes.com/2011/07/28/advanced-parallax-propeller-robot-big-brain/
Dangerous Prototypes showcases a new open source hardware project every month.
Modern Mechanix April 1935
“Scientists Invent Machine To Discover How Brain Works
THE brain, perhaps the most mystifying organ of the human body, can now be scientifically studied by a new apparatus which photographs amplified “action currents.” Invented by Dr. H. H. Jasper and Dr. L. Carmichael of Brown University, the new machine will permit physicians to study the action of the brain just as the electrocardiograph permits a revealing study of heart action.
A headpiece on the head of the patient picks up electric currents of about one ten-millionths of a volt which flow from the brain in waves, at a rate of from eight to fifty per second. The currents are carried to an amplifying box where they are intensified 500,000 times and flashed across a glass disc. The ordinary currents are smooth and wavy; when the mind is disturbed, they are sharp and irregular.”
http://blog.modernmechanix.com/2009/10/28/scientists-invent-machine-to-discover-how-brain-works/
It's interesting how the same underlying principle of technology is used today in reading brain waves as it was in 1935, over seven decades ago.
Another Propeller is added to the collective
This powerful Propeller system adds useful resources to the Big Brain and can serve as the Left Brain's Front End or Rear End.
http://www.parallax.com/Store/Microcontrollers/PropellerKits/tabid/144/CategoryID/20/List/0/SortField/0/Level/a/ProductID/467/Default.aspx
Here’s what the Big Brain picked up by assimilating HYDRA
HYDRA Development Board Console with 128K EEPROM and I/O interfaces
Propeller CPU at 160 MIPS, (80 MHz, 20 MIPS per cog) 8 PIC-like CPU cores each with 496 32bit-words of workspace plus a shared 32kB RAM and 32kB ROM
128kB serial EEPROM Re-Programmable Storage Card
two NES Compatible Game Ports
one USB Programming port
one Serial Port or second USB port
two PS/2-compliant Ports for Mouse and Keyboard.
one NTSC or PAL Video Output, or alternatively one VGA Video Output
one mono Audio output
one 20-Pin Expansion Interface connector
one RJ-11 Peer-to-Peer Networking Port
HYDRA Manual - "Game Programming for the Propeller Powered HYDRA" hard copy book by Andre' LaMothe
Blank "Experimenter" Card to Design Your own Add-On Hardware
PS/2 Optical 3-button Mouse with scroll
88 QWERTY PS/2 Mini Keyboard
Nintendo Compatible Gamepad
A/V Cable
USB Programming Cable
CD-ROM with all source, demos, and development tools
Video
As the video display is generated in software, the capabilities of it depend on the software driver, but a typical PAL/NTSC driver would be able to generate a 256x192 screen resolution, and typically 80 to 96 colors maximum [1]. The color resolution for the VGA display is hardwired to 64 colors maximum, and a typical resolution would be 800x600 depending on the available display memory. A video driver uses the main 32kB RAM which it must share with code space. With the aid of the new 512kB memory expander, video drivers can be written that extend the video generator's capabilities. To aid in the display of text the Propeller chip has a complete character font in its ROM.
The HYDRA Game Console is based on a socketed 40-Pin dip version of the new Parallax Multiprocessing Propeller Chip. The Propeller chip is simplified cell processor very similar in concept to the Sony "Cell Processor" used in the Playstation 3. The Propeller chip is the main CPU of the HYDRA and runs at 80 MHz for an overall processing power of 160 MIPS (20 MIPS per processor). The Propeller has 32K of internal static RAM used for program storage and work space. While the HYDRA Game Console has a 128K of EEPROM storage. The HYDRA supports numerous interface devices such as NTSC/PAL video, VGA video, PS/2 ports, networking, Nintendo compatible gamepads as well as a 20-pin expansion port header used to plug small 1x2.5" cards into. Also, as Propeller chips are upgraded the 40-Pin dip Propeller chip can be removed from the HYDRA and upgraded since its socketed. Additionally, the main clock crystal on the HYDRA can be replaced for "overclocking experiments".
The HYDRA Graphics and Sound
The HYDRA graphics and sound are generated by the Propeller chip's internal hardware via a hardware/software combination. The HYDRA is capable of generating very high resolution NTSC and VGA displays with up to 256 color variations. The basic palette of the HYDRA is 86 colors. Screen resolutions are 100% software programmable and thus the HYDRA can generate almost any kind of tiled, bitmapped, or sprite based displays with the proper programming. Numerous games and graphics demos show off the different techniques to generate graphics on the system. Sound on the HYDRA is generated via a 1-bit PWM (pulse width modulation) system. Currently, drivers come with the HYDRA that can support up to 11 channels with 22 KHz playback of pure tones as well as digital samples. A complete ADSR (Attack-Sustain-Decay-Release) software engine is included. Users of the chips like the AY8910 and SID chips from the C64, Atari, and other 8-bit computers will feel right at home.
The Propeller Chip
The Propeller chip is a 32-bit symmetrical multiprocessor with 8-independent cores. Each core can run its own program and has its own data. Thus, the Propeller is a MIMD (Multiple-Instruction-Multiple-Data) system. Basically, its like having 8 processors on the HYDRA. Each core called a "COG" has a local memory of 512 32-bit words that are used for both instructions and data. Additionally, each COG has 2 counters with numerous modes of operations, PLLs, and video serializing hardware. Thus, each COG is actually capable of generating video independently. The only coupling between the COGs is thru a common shared 32K Byte memory. Each COG accesses the shared memory on a "round robin" schedule. The architecture is RISC based with usually 4 clocks per instruction.
Programming and the Propeller IDE
Programming the HYDRA Game Console's Propeller chip is done via the Propeller Integrated Developed Environment (IDE). There are two languages that can be used: SPIN and assembly language. SPIN is very similar to C/BASIC and for beginners, while assembly language programming is for those that want the ultimate in control and speed. In either case, both languages are integrated seamlessly in the IDE and you simply write your code, compile, and download to the HYDRA via a USB connection (USB 1.0 or above required, cable included).
Downloads & Resources at
http://www.parallax.com/Store/Microcontrollers/PropellerKits/tabid/144/CategoryID/20/List/0/SortField/0/Level/a/ProductID/467/Default.aspx
HYDRA Demo Source Code (.zip)
HYDRA Demo Video (.wmv)
The Full HYDRA Manual (.pdf)
The Hydra CD and Resources (ftp)
Design your own Computer Games for the Propeller Chip (.pdf)
3D Game - Michael Park (.wmv)
Circuit Cellar/Game On by Tom Cantrell (Off Site)
Mercury News Blog Coverage of the HYDRA Kit (Off Site)
Mercury News Article on the HYDRA Game Developing Kit (Off Site)
XGameStation Web Site (Off Site)
Acrylic enclosure box for this board from Mountain King (Off Site)
Links
http://en.wikipedia.org/wiki/HYDRA_Game_Development_Kit
http://www.xgamestation.com/browse_products.php?category=7
http://www.xgamestation.com/view_product.php?id=33
http://www.xgamestation.com/products/hydra/hydra_ss_01_large.jpg
http://www.avalondreams.com/hydra/index.html
http://www.xgamestation.com/media/hydra_poster_parallax_logo_01.jpg
http://www.savagecircuits.com/forums/showthread.php?393-Hydra-Gaming
http://forums.parallax.com/showthread.php?108737-Xmas-gift-for-new-users-Propeller-Software-Index-updated-12-20-08&highlight=christmas+hydra
Propeller Software Index
http://forums.parallax.com/attachment.php?attachmentid=57451&d=1229778213
http://forums.parallax.com/showthread.php?111166-Propeller-HYDRA-Key-Thread-Index
http://forums.parallax.com/showthread.php?108626-HYDRA-Projects-Master-List
I was thinking it should have at least one arm with a feedback control loop for each side, and perhaps some swarm-bots sending in environment data (at least something simple, like temperature, light level, etc).
Is this already started?
prof_braino, interesting you should mention that. I was researching swarm bots today. It's an ongoing project. The Big Brain with many processors is perfect for controlling a swarm of robots. My interest is in a top mount airport pad. In the evening I reviewed some hobby helicopter crafts being flown at the community park.
From the brain airport, tiny copters could launch for reconnoissance missions etc. I'm looking into some of the smallest commercially available air copter products. Some have very tiny cameras run by a button cell. Another idea, I have enough Penguin robots to do a ground swarming.
Another Big Brain sub-project is started for sensors. There's plenty of sensors to wire up but with so much power, the application must be carefully chosen. Currently this is in a stall trying to decide the best applications and ones simple enough that won't take a year to complete.
The sensor I/O devices can include some stuff in my junk box and from other projects - dual and tri-axis accelerometers, PIR detector, IR transmitter, IR Receiver, cadmium photocell, phototransistor, radio transmitter, radio receiver, Vibra Tab Mass, GPS, PING))), digital thermometer, pressure, humidity, gas, compass, sound in, sound out, speech in, speech out, LCDs, TV, keyboard, keypad, coprocessors, Hall Effect magnetic sensor, data logger, light to frequency converter, piezo speaker, small solar cell, amplifier, a sound converter, a mobile robot transport, etc.
Erco:
The arm is already started. The name is Brain Stub changed to BASIC Stump. Remember? You coined that one! A search on the names should bring up some preliminary postings with details considered.
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=1019628&viewfull=1#post1019628
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=1019843&viewfull=1#post1019843
http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain&p=1015760&viewfull=1#post1015760
This project may divert from the large massive arm on the big 5-foot rack to a smaller stump that mounts directly on the Left Brain's smaller EXOskeleton, or include both. The smaller stump would make a more portable section to the Brain that could move around on its own and transport more easy. I have all the needed parts and materials for this project.
Talk with Right Brain in Four Languages
It's now possible to carry on a conversation with the Right Brain in your choice of several languages, with this AI program accessed online.
Tom Riddle's Diary is a chatterbox that speaks four languages - English, German, French and Italian. For the first time, you can now speak with the Right Brain in multi languages.
http://pandorabots.com/pandora/talk?botid=c96f911b3e35f9e1
This is a wireless multi-lingual add on function and is the beginning of experimental features provided by internet and accessed by the Big Brain.
TRD is simply online linked to the Right Brain giving it the ability to converse with you using its online database. TRD is one of the most popular Pandorabots.
Full List of Pandorabots
http://www.pandorabots.com/botmaster/en/~1b3734620bb83f30400c3236~/mostactive
Other Links
http://www.chat-bot.com/chatbots.php
http://www.google.com/#hl=en&cp=6&gs_id=l&xhr=t&q=CHAT+BOT&pf=p&sclient=psy&source=hp&pbx=1&oq=CHAT+B&aq=0&aqi=g4g-s1&aql=&gs_sm=&gs_upl=&bav=on.2,or.r_gc.r_pw.&fp=d97a493102ffbda7&biw=942&bih=619