Computer science

The registered models of skcomplex.models.computer_science. The page of each shows its factory, its component types with their states, parameters, and laws as written, and, for the forms that have them, its equations as SymbolicAnalysis returns them, so it serves as the pattern for a model of your own.

Automata

• Automata on the Grigorchuk group: The majority automaton on a ball of the Grigorchuk group, a group of intermediate growth.
• Automata on the free group: The majority automaton on a ball of the free group, a tree in which every cell follows its neighbours.
• Baumslag-Solitar groups: The majority automaton on a ball of the Baumslag-Solitar group BS(m, n), of exponential growth.
• Diffusion automaton: A cellular automaton of diffusion, c' = c + r Σ (c_neighbour - c), started from a square of ones.
• Elementary automata: An elementary cellular automaton by its Wolfram number, started from a single live cell by default.
• Game of Life: Conway's Game of Life, or another life-like rule, on a torus.
• Heisenberg: The majority automaton on a ball of the discrete Heisenberg group, of polynomial growth of degree 4.
• Highlife and the planar automaton zoo: HighLife, B36/S23, or any other life-like automaton of the planar zoo by its rule, on a torus.
• Lamplighter: The parity automaton on a ball of the lamplighter group, a linear rule over the field of two elements.
• Langton's ant: Langton's ant, which wanders chaotically for about 10,000 steps before it builds a periodic highway.
• Linear automata over rings: The discrete Laplacian as a linear automaton on a ring, over the field of modulus elements.
• Noisy automata: An elementary cellular automaton whose outputs are flipped at random with a small probability.
• Subshifts: A Markov chain on the golden mean shift, the configurations of a ring that never hold two adjacent ones.
• Unbounded-alphabet automaton: An additive automaton over the nonnegative integers, x'(n) = x(n - 1) + x(n), from a single 1.

Evolutionary computation

• Bit-string evolution on a fitness landscape: A population of bit strings that evolves on a random additive fitness landscape.
• RobbyGA can collector: Mitchell's genetic algorithm that evolves strategies for Robby, a robot that collects cans in a room.
• Simulated annealing, genetic algorithms, and differential evolution: A population of candidates that searches a two-dimensional landscape, one generation per step.

Neural networks

• Asymmetric diluted Hopfield network: The asymmetric diluted Hopfield network, which recalls a stored pattern below its memory capacity.
• Binary perceptron and random K-SAT: A binary perceptron that must store random patterns, revealed one per step, as cardinality constraints.
• Echo state network: An echo state network, a reservoir of leaky tanh units with random recurrent weights driven by a sine.
• Elman recurrent network trained by backpropagation: Elman's simple recurrent network, which learns to predict the next bit of a temporal XOR stream.
• Hopfield network: A Hopfield network that stores patterns with Hebb's rule and recalls the first from a corrupted copy.