Get Updates By E-mail 

Article Written By:
Johann Dowa

AIToolBox – Swift Based Library Providing An Extension Collection Of AI Helper Modules

  •  
  •  
  •  
  •  
  •  

AIToolBox is an open source library Swift library from Kevin Coble providing a multitude of AI modules.

The modules include genetic algorithm helpers, graphs/trees, alpha-beta, neural networks and more. The Accelerate framework is used to speed up computations.

This listing from the readme shows the different algorithms included:

Graphs/Trees
    Depth-first search
    Breadth-first search
    Hill-climb search
    Beam Search
    Optimal Path search

Alpha-Beta (game tree)

Genetic Algorithms
    mutations
    mating
    integer/double alleles

Constraint Propogation
    i.e. 3-color map problem

Linear Regression
    arbitrary function in model
    convenience constructor for standard polygons
    Least-squares error

Neural Networks
    multiple layers, several non-linearity models
    on-line and batch training
    simple network training using GPU via Apple's Metal

Support Vector Machine
    Classification
    Regression
    More-than-2 classes classification

K-Means
    unlabelled data grouping

Principal Component Analysis
    data dimension reduction

Markov Decision Process
    value iteration
    policy iteration
    fitted value iteration for continuous state MDPs - uses Linear Regression class for fit
            (see my MDPRobot project on github for an example use)

Gaussians
    Single variable
    Multivariate - with full covariance matrix or diagonal only

Mixture Of Gaussians
    Learn density function of a mixture of gaussians from data
    EM algorithm to converge model with data

You can find AIToolbox on Github here.

A nice collection of AI helpers.


  •  
  •  
  •  
  •  
  •  

Submit A Resource

Have you created a useful tutorial, library or tool for iOS development that you would like to get in front of our 300,000+ monthly page views from iOS developers?

You can submit the url here.

The resources we feel will appeal to our readers the most will be posted on the front page.

Comments