App Detail » Neuro Net

Published by: Mic-Apps Limited

iPad App - Designed for iPad

Price: $3.99
Current Version: 3
Released: October 27, 2012

What's New

General refresh update after iOS 10

App Description

Neuro Net is the only Neural Network app available on the app store !

With Neuro Net, you have in the palm of your hand a very powerful and fully ready-to-be parametrized neural network.

SIMPLE & "EASY TO USE" --------------------
- Use the efficient touch interface to define your training function
- Define the property of your neural network
- Use an advanced genetic algorithm to train your network
- Use robust back propagation to fine tune your solution
- Neuro Net is a useful tool and should be easy enough to use even for novice however prior knowledge of neural networks would greatly help in understanding all the parameters ! Several help pages are provided.

BESPOKE NEURAL NETWORK --------------------
- Define the nature and complexity of your own network
- Choose the number of input parameters (up to 20 input neurons)
- Choose the number of hidden layers (up to 4 layers)
- Choose the number of neurones on each hidden layers (up to 20 neurons each)
- Choose the type of activation function
- Choose thresholds

FULL CONTROL --------------------
- Fine tune settings of the genetic algorithm
- Fine tune settings of the back propagation method
- Combine both genetic algorithm and back propagation for best results

EXPLORE NETWORK --------------------
- Once neural network is trained, use the tools to explore output to any value
- Explore the network properties: weight, bias, threshold, activation type

PROJECT MANAGEMENT --------------------
- An unlimited number of projects can be saved and loaded with any names
- Genetic algorithm, back propagation and network settings can be saved and recalled independently

EXAMPLES --------------------
- Neuro Net comes with several examples
- Examples include training set, solution and settings

APPLICATIONS --------------------
Neuro Net is the best way to learn about neural networks and to build powerful multi-input, single-output networks for many applications, including:
- Predict or extrapolate suite of numbers
- Optimisation of complex functions
- Replicate complex and time consuming simulations
- Use network to find optimum or explore design domains