Download

We need your help! 🔥

Please consider disabling your ad-blocker to support this website! 🔑
We are a free, community-supported website! 🥰
We only display small unobtrusive ads which help us stay online! 💯
Thank you in advance! ❤️


Evolutionary Learning


Explicit Components

Loop

Octopus


Supervised Learning

Utilities

View

Octopus

ADDON. Version 0.4. Released on 2018-Dec-05. Provides 35 components. Created by Robert Vierlinger. Features 0 video tutorials.

Octopus was originally made for Multi-Objective Evolutionary Optimization. It allows the search for many goals at once, producing a range of optimized trade-off solutions between the extremes of each goal. It is used and works similar to David Rutten's Galapagos, but introduces the Pareto-Principle for Multiple Goals.

Explicit Components

	Check Pareto Dominance Takes two individuals and returns the dominance relation, assuming a minimization problem.
	Compute Diversity Computes the paremetric diversity of each solution in the list and adds it as an objective
	Construct Solution Takes a lists of numbers for parameters and/or objectives to wrap them in a single object for better handling of pools of solutions [generations etc.]
	Cull Duplicate Solutions Removes duplicate solutions
	Deconstruct Solution Takes an OctopusSolution object and explodes it into parameters and objectives
	Cull Elite Selects an elite of best multi-objective solutions, assuming a minimization problem.
	Cull Pareto Fronts (Pareto Fronts) Divides a set of solutions into pareto-fronts, assuming a minimization problem.
	Hypervolume Contributions (HV+) Calculates the Hypervolume contributions of a multi-dimensional set of points in relation to a reference point
	Hypervolume Calculates the Hypervolume of the ParetoFront of a multi-dimensional set of points; exact algorithm; normalizes the pareto front to objectives between 0 and 1
	Mutate (Mutate an Octopus Solution) Mutate a solution's parameter values
	Remap Objectives Remaps the objective values of a set of solutions, assuming a minimization problem when taking the pareto fron as a start domain.
	Crossover (Simulated Binary Crossover) Takes two individuals and exchanges parameters between them - after 'SBX - Simulated Binary Crossover'
	Tournament Selection (TS) Tournament Selection for single or multi objective solutions, assuming a minimization problem.

Evolutionary Learning

	Breeder Settings - All (oEL) (SetAll (octEvoLearn)) Settings for the NEAT Algorithm to evolve an ANN
	Breeder Settings - Basics (oEL) (SetBase (octEvoLearn)) Basic Settings for the NEAT Algorithm to evolve an ANN. These will override any changes made to the properties in the All-Settings Component.
	Construct Network (oEL) (ConsN (octEvoLearn)) Create a Network by node-points, connection-indices and weights.
	Deconstruct Network (oEL) (DeconN (octEvoLearn)) Deconstruct a network into its nodes, connections, weights, functions and metadata like performance data.
	Deconstruct Network Obj (oEL) (DeconNObj (octEvoLearn)) Gives the Network's Objective and Fitness Values
	Evaluate Network (oEL) (EvalN (octEvoLearn)) Forward-Pass through the Network: Takes values for the input nodes and calculates the outputs.
	Field Curve (oEL) (FC (octEvoLearn)) Draws a curve following the direction field defined by a Network. Integration with Runge-Kutta 4th order.
	Modify Weights (oEL) (Crossover (octEvoLearn)) Takes a pool of Networks and produces offspring by crossover mating
	Mutate Weights (oEL) (MutW (octEvoLearn)) Mutate connection weights of a Network
	Breeder (oEL) (NEAT (octEvoLearn)) Evolves artificial neural networks with the NEAT algorithm, using SharpNeatLib by Sebastian Risi
	Random Network (oEL) (RN (octEvoLearn)) Generate a random network
	Show Network (oEL) (SN (octEvoLearn)) Opens a window to show the Network

Supervised Learning

	Network Training Settings (oSL) (NTS (octSupervLearn)) Settings for RPROP Supervised Learning of an ANN
	Network Evaluate (oSL) (EvalN (octSupervLearn)) Evaluate a network for some input values
	Network Learning (oSL) (NetLearn (octSupervLearn)) Supervised example training of a Network by multi-core resilient propagation algorithm, using the Encog library by Jeff Heaton
	SVM Evaluate (oSL) (SVMEval (octSupervLearn)) Evaluate the learnt SVM function
	SVM Learning (oSL) (SVMLearn (octSupervLearn)) Train SVM and optionally estimate parameters using grid search and cross validation

Loop

	Octopus Loop Octopus Loop
	Octopus Evaluator (Octopus Eval) Octopus Evaluator

Utilities

MD NearestNeighbors (MD NearN)
Neighbourhood search of multi-dimensional points (euclidean kd-tree)

View

Select Solutions
Takes octopus solutions or networks with saved phenotype meshes to show them for selection

Octopus

Octopus
Multi-objective seach and optimzation