This neural network has better handwriting than you
Alex Graves from the University of Toronto has engineered a neural network which can generate handwritten notes from typed text samples.
In a paper, Graves explains the mechanism for the program. The network learns by tracing the placement of pen tips as people write, rather than taking written samples of existing handwriting.
You can generate your own samples here.
FACTS ABOUT NEURAL NETWORKS
Artificial neural networks are intended to mimic the biological neural networks of the human brain. The brain is estimated to have about ten to one hundred billion neurons while artificial networks generally have about 1,000.
Neural networks can compute any equation, regardless of the amount or complexity of input and output points.
Knowledge in neural networks is distributed throughout the program instead of being coded into it. Neural networks have been used to detect handwriting, predict bombings, and make investment decisions.
Cover image: University of Toronto