NeuroSolutions for MATLAB - 神经网络工具箱

NeuroSolutions for MATLAB神经网络工具箱是MATLAB的技术计算功能的有用的附加工具，其允许用户在MATLAB中使用NeuroSolutions的功能。该工具箱将16种神经模型，5种学习算法和大量的应用在易用的用户界面中，几乎对神经网络无了解的用户都可以使用该产品。该工具箱也与NeuroSolutions进行了，使得用户可以在NeuroSolutions中创建自定义的神经网络，并通过NeuroSolutions for MATLAB的用户界面在MATLAB中对其进行使用。

 

神经网络工具箱是MATLAB，外接程序，使用方便，直观。它利用了业界的神经解决方案、神经网络和MATLAB中的人工智能。该软件可以让你集中精力解决你的问题，而花很多钱去寻找神经网络文献和自己开发算法。

 

易于使用的

MATLAB的神经解可以被“几乎不了解”神经网络的用户所利用。熟悉MATLAB的用户可以在几分钟内使用整个软件包。您可以将NeuroSolutions用于MATLAB的数据拟合、模式识别、时间序列预测等应用程序。

 

的引擎

NeuroSolutions for MALTAB的特点是拥有NeuroSolutions的神经网络引擎。23位和64位版本的MATLAB这是在MATLAB中使用神经网络的的方法。对于较大的网络，使用NVIDIA CUDA™或OpenCL™与传统的CPU相比，使使培训时间从小时提高到分钟。

 

神经解决方案的MATLAB套件：

NeuroSolutions for MATLAB也通过NeuroSolutions for MATLAB NeuroSolutions，该套件使用用户能够在NeuroSolutions中构建自定义网络，使用自定义解决方案向导为这些网络生成基于Windows的DLL，然后使用NeuroSolutions for MATLAB 中的DLL。

 

工具箱有16个神经网络结构，7个学习算法和一大堆有用的实用工具，使人们能够利用神经网络的力量来解决复杂的现实问题。

 

该工具箱以下神经网络的变形：

• 多层感知机

• 通用前馈神经网络

• 概率神经网络

• 模块化神经网络

• 向量机

• 部分循环神经网络

• 循环神经网络

• 延迟循环神经网络

 

除了Levenberg-Marquardt方法以外，其还提供了以下6种学习算法

• 步骤法

• 动力法

• Quickprop法

• Delta-Bar-delta法

• 共轭梯度法

• Levenberg-Marquardt法

 

工具箱还一些实用工具，符号数据转换，允许使用文本数据作为神经网络的输入。符号数据转换函数与反向转换函数一起使用，后者在完成任务后将数据转换回符号。图像扁平化工具将图像扁平化成一行数据，这样可以将其输入神经网络。性能指标函数揭示了神经网络与统计指标的训练效果。

 

64位计算

提高性能与Windows 64位计算和的可执行代码显著缩短训练时间！

 

有弹性的反向传播（RProp）：

弹性反向传播是前馈人工神经网络中监督学习的一种启发式学习学法。与Levenberg-Marquardt相比，弹性反向传播是现有快的全职更新算法之一。

 

R2013和64位MATLAB的神经方案已经被重写，以更地使用MATLAB的新框架，对R2013/32位和64位安装。

 

NeuroSolutions for MATLAB更新至V4，更新的内容如下：

● 提升了在Windows 64位系统上的计算性能，了可执行代码，培训时间显著缩短。

● 弹性反向传播是前馈式人工神经网络中监督式学习的一张启发式的学习方法，次于夸特法（Levenberg-Marquardt）。弹性反向传播是目前快的权值更新算法之一。

● 在使用夸特法网络中将NeuroSolutions加速器扩展组件与NVIDIA™和OpenCL™显卡或处理器一起使用，训练时间从数小时提高到短短几分钟—这是反向传播学习有力的一种形式。

● NeuroSolutions for MATLAB的大部分已经重写使能以方式与MATLAB新的架构一起工作，R2013和32/64位的安装装置。

 

信息

MATLAB版本：

2008、2009、2010、2011和2013（的版本为32和64位）

 

 

【英文介绍】

NeuroSolutions for MATLAB neural network toolbox is a MATLAB；add-in that is easy-to-use and intuitive. It leverages the industry leading power of NeuroSolutions neural networks and artificial intelligence inside MATLAB. The software allows you to concentrate on solving your problem without having to spend many hours persuing neural network literature and developing the algorithms yourself.

 

Easy to Use

NeuroSolutions for MATLAB can be utilized by users with "next to no knowledge" of neural networks. Users who are familiar with MATLAB will be able to jump in and use the entire package within a few minutes. You can use NeuroSolutions for MATLAB for applications such as data fitting, pattern recognition, time-series prediction and much more.

 

Powerful Engine

NeuroSolutions for MALTAB features the neural network engine from industry leading NeuroSolutions. Supporting both 32- and 64-bit versions of MATLAB it is the most powerful way to use neural networks in MATLAB. For larger networks, use NVIDIA CUDA™ or OpenCL™ to enable training time improvements from hours to minutes when compared to traditinal CPU's.

     

NeuroSolutions for MATLAB Suite

NeuroSolutions for MATLAB is also integrated with NeuroSolutions through the NeuroSolutions for MATLAB Suite that enables users to build custom networks in NeuroSolutions, generate Windows-based DLL's for those networks using the Custom Solution Wizard and then use the DLL's inside NeuroSolutions for MATLAB.

 

The toolbox features 16 neural network architectures, 7 learning algorithms and a host of useful utilities that enables one to employ the power of neural networks to solve complicated real-world problems.

 

Neural Network Architectures:

Multilayer Perceptron

Generalized Feed Forward Network

Probabilistic Neural Network

Modular Neural Network

Support Vector Machine

Partially Recurrent Neural Network

Fully Recurrent Neural Network

 Time-Lag Recurrent Neural Network

 

Learning Algorithms:

Step

Momentum

Resilent Backpropagation (RProp)

Quickprop

Delta-Bar-Delta

Conjugate Gradient

Levenberg-Marquardt

 

The toolbox also includes several utilities including Symbolic Data Translation allows for using textual data as inputs to a neural network. The symbolic data translation function comes paired with the inverse translation function that translates data back to symbols once all other tasks are completed. The Image Flattening utility flattens an image into a single row of data, so that it can be fed into a neural network. The Performance Indicators function reveals how well the neural network has trained with statistical indicators.

 

64-bit computing

Increase performance with Windows 64-bit computing and optimized executable code for significantly shorter training times!

 

Resilient Backpropagation (RProp)

Resilient Backpropagation is a learning heuristic for supervised learning in feed forward artifical neural networks. Next to Levenberg-Marquardt, Resilient Backpropagation is one of the fastest weight updating algorithms available.

 

NeuroSolutions Accelerator Add-on

See training time improvements from hours to minutes using NVIDIA™ and OpenCL™ graphics cards or processors with the NeuroSolutions Accelerator add-on on networks using Levenberg-Marquardt - the most powerful form of back-propagation learning available.

 

Supporting R2013 and 64-bit MATLAB

Much of NeuroSolutions for MATLAB has been rewritten to work with MATLAB's latest framework more optimally including support up to R2013 and 32- and 64-bit installations.

 

Getting Started with NeuroSolutions for MATLAB

Visit our Video Library page for more videos on NeuroSolutions and related products.

Other Information

 

Supported MATLAB Versions:

2008, 2009, 2010, 2011, 2012 & 2013 (32/64-bit for all supported versions)

Still unsure about NeuroSolutions for MATLAB?