Herself's Artificial Intelligence

Humans, meet your replacements.

Braitenberg Vehicles




Synthetic psychology is a field where biological behavior is synthesized rather than analyzed. The father of such behavior Valentino Braitenberg ( home page ) did some interesting work with toy vehicles in the 1986 book “Vehicles: Experiments in Synthetic Psychology“.

The Braitenberg vehicles were simple toy cars with light sensors as headlights. In some cars the wire for each headlight went to the real wheel directly behind it, in some the wires went to the opposite back wheel. The headlight receptors were aimed slightly off to the outside. The more light received by a receptor the faster the wheel wired to that receptor would turn.

Each vehicle exhibited realistic behavior when placed in a plane with several randomly placed light sources. A vehicle wired straight back when placed near a light source will then rush towards the light veering off as it gets closer to the light. As it gets more distant from the light sources the vehicle slows down. The reason is the wheel receiving the most light spins fastest turning the car away from the light source.

The vehicle with the crossed wires will turn and drive towards the brightest light in its field of view. The closer it gets, the faster it goes eventually running into the light.

Pretty interesting behavior from a very simple machine. But what if we add in a neurode to each wire and instead of a plain wire we use a wire that inhibits signals? Neurodes are set to only fire if they receive signals over a certain threshold. In this case zero is to be our threshold. So now our cars send signals to the wheels if there is no light, and do not send a signal if there is a light. Now the vehicle with the wires straight back moves toward a light and slows as it approaches, facing the light. The second vehicle now avoids light sources, speeding off to the darkest corner it can find.

Using thresholds ( see logic version ) you begin to see group behavior, one set of vehicles herds and traps another set.

So what has this all to do with current artificial intelligence? Some of our best stuff right now came from earlier work that was done and stopped. Some of our best mathematical algorithms come from extremely early math. And to remind you ( and me ) that simple rules can create very complex behavior in game characters and artificial life forms.

Four neurode versions of these vehicles have been built and they will exhibit more complex, non-linear behavior. “The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.” (Edsger Dijkstra) At what point does the behavior become realist enough to be considered a life form?

Source Code:
ObjC simulation of Braitenberg Vehicles( #2 fear and aggression )
ObjC simulation of Braitenberg Vehicles( #3 love, explore and complex )
ObjC simulation of Braitenberg Vehicles( #4 values and tastes )
ObjC simulation of Braitenberg Vehicles( #5 Logic )

More information:
Braitenberg Vehicles ( Java and Lisp simulators here )

Papers:
Swarm modelling. The use of Swarm Intelligence to generate architectural form. ( pdf)

Braitenberg vehicles Fear and Aggression

Braitenberg vehicles
Fear and Aggression


SantaFe Ant Trail




I’m revisiting this problem again and hoping to dig more into evolutionary computing.

The SantaFe Ants Trail is a maze of 89 food pellets, with breaks in between. The ants must find a way to pick up all of the pellets.

Each ant checks to see if there is food in front, moves forward and collects it if so, else turns left, right or moves one forward depending on its coding. The food is removed for this ant once it is collected.

The genome must be a minimum of 32 chromosomes, that’s the minimum the state machine must have to solve the maze. It’s been solved with 32 chromosomes, 200 steps, 200 generations, 65,536 starting ants.

The fitness of the ant is the number of pellets it has collected.

Each generation the top 50% of the ants mate, half the genes ( in sequence from one parent, half from the other )

I only released 2048 ants at once, I suppose with out graphics the computer would’ve handled much more but it was interesting to watch and I spotted several problems along the way while watching.

At 512 ants, my best score was 43, at 1024 best was 50.

Problems I ran into:
- If one or a few ants were far above rest they cloned the entire population and diversity was lost.
- Need a way to cull individuals too far from the center and or remove clones. I inserted random mutations rather than check for clones, sorting by genes might make it easy to id and remove clones.

Source code:
ObjC iOS SantaFe Ant Trail

More info:
Wiki SantaFe Ant Trail
Solving the Artificial Ant on the Santa Fe Trail Problem in 20,696 Fitness Evaluations

See also:
Evolutionary AI

News:
Artificial Intelligence: Riders on a Swarm


DeepBeliefSDK




This is very cool, iOS video image recognition.

I am totally convinced that deep learning approaches to hard AI are going to change our world, especially when they’re running on cheap networked devices scattered everywhere. I’m a believer because I’ve seen how good the results can be on image recognition, but I understand why so many experienced engineers are skeptical. It sounds too good to be true, and we’ve all been let down by AI promises in the past.

That’s why I’ve decided to release DeepBeliefSDK, an iOS version of the deep learning approach that has taken the computer vision world by storm. In technical terms it’s a framework that implements the full Krizhevsky stack of 60 million neural network connections, with a customizable top layer inspired by the Decaf approach. It does all this in under 300ms on an iPhone 5S, and in less than 20MB of memory. Here’s a video of me of me using the sample app to detect our cat!
More at Pete Warden’s Blog

GitHub, DeepBeliefSDK


Java Neural Network Framework Neuroph




Neuroph is lightweight Java neural network framework to develop common neural network architectures. It contains well designed, open source Java library with small number of basic classes which correspond to basic NN concepts. Also has nice GUI neural network editor to quickly create Java neural network components. It has been released as open source under the Apache 2.0 license, and it’s FREE for you to use it.

Neuroph simplifies the development of neural networks by providing Java neural network library and GUI tool that supports creating, training and saving neural networks.

If you are beginner with neural networks, and you just want to try how they work without going into complicated theory and implementation, or you need them quickly for your research project the Neuroph is good choice for you. It is small, well documented, easy to use, and very flexible neural network framework.

Neuroph


Top online education resources for math, artificial intelligence and machine learning




The number of online education opportunities continues to grow. I’ll post them here as I hear about them.

Firm start/stop dates and certificates:
EdX
Coursera
Udacity

Firm start/stop date, no certificates
Stanford Online

On your own schedule – no certificates:
Complexity Explorer – Santa Fe Institute
Video Lectures (Machine Learning)
Khan Academy
iTunesU
MIT Open Courseware

Right now computer science, including machine learning, artificial intelligence and math make up most of the classes.