Skip to main content

neural network 

A way of doing AI.
It's rally cool.
You can teach it anything.
If you give it enough
time, information and space
it would be as smart as you
(or eaven smarter...)
Neural Networks rules

recurrent neural network

A recurrent neural network is a feedforward neural network with short term memory.
Programmer 1: Have you upgraded your neural network to a recurrent neural network?
Programmer 2: Yes, it's no longer a feedforward neural network.
Programmer 3: Recurrent neural network? Why not a long short term memory?
Programmer 2: I'll upgrade it again.

Creative Neural Networks

Neural network architectures designed not just for pattern recognition and prediction but for genuine creativity—generating novel outputs that weren't implicit in training data. Creative Neural Networks would combine the pattern-matching power of deep learning with mechanisms for genuine novelty: constrained randomness, cross-domain synthesis, emergent recombination. They wouldn't just optimize for accuracy; they'd optimize for interestingness, for novelty, for surprise. The frontier where neural networks stop recognizing and start imagining.
"The neural network didn't just classify images; it generated new species of flowers that looked like they could exist but didn't—combining features across genera in ways no botanist had seen. That's Creative Neural Networks: not recognizing the world, but expanding it."