ANNHUB

Machine Learning Studio

Design, train, and validate custom machine learning models with a few simple clicks. No programming skills are required.

Resources

Free Trial

How It Works

ANNHUB supports datasets in comma-separated values (CSV) format. The outputs are identified by keywords "output, target, class." Each row in the csv file is equivalent to a data sample.

ANNHUB supports various activation types for both hidden layer and output layer. Both pre-processing and post-processing methods are provided to normalize dataset inputs and outputs. Designers can use different cost functions, together with training parameters to customize a neural network architecture that suits their applications. ANNHUB also autosuggests the suitable neural network architecture based on input data.

ANNHUB provides evaluation techniques such as ROC curves, confusion matrix, and regression curves to evaluate the performance of the trained neural network on both the training set, validation set, and test set. These ROC curves provide essential information that helps the machine learning design process.

Users can load a completely new dataset into ANNHUB to evaluate the performance of the trained model before deciding to deploy it into a real-time application.

Product Features

No coding

ANNHUB is specifically designed for engineers who do not have a profound knowledge of machine learning and programming skills.

ANNHUB is easy to use and has various features to help you create accurate and reliable machine-learning models. These features include:

A drag-and-drop interface that makes it easy to create and modify your models
A built-in algorithm advisor that recommends the best algorithms for your data set
A visualization tool that helps you understand the performance of your models

Advanced algorithms

ANNHUB supports a variety of advanced algorithms, including:

Scaled Conjugate Gradient (SCG): A gradient-based optimization algorithm that is more efficient than traditional gradient descent algorithms.
Levenberg Marquardt (LM): A hybrid algorithm that combines the advantages of gradient descent and Gauss-Newton methods.
Quasi-Newton: A family of algorithms that use an approximation to the Hessian matrix to improve the convergence rate of gradient descent.
Bayesian Regularization: A method of regularizing machine learning models by adding a penalty term to the loss function that penalizes the complexity of the model.

These algorithms can be used to optimize the training speed and convergence rate of machine learning models.

Design assistant

ANNHUB can auto-suggest neural network architectures for performance optimization. This means that you can simply upload your data set, and ANNHUB will recommend the best neural network architecture for your specific task. This can save you a lot of time and effort in designing an AI model.

Flexible model testing

ANNHUB supports a variety of model evaluation metrics, including:

ROC curves: A graphical plot of the true positive rate (TPR) against the false positive rate (FPR). ROC curves are used to evaluate the performance of binary classification models.
Confusion matrix: A table that summarizes the performance of a classification model. The confusion matrix shows the number of true positives, false positives, true negatives, and false negatives.
Performance metrics: A set of metrics that quantify the performance of a machine learning model. Common performance metrics include accuracy, precision, recall, and F1 score.
Regression curves: A graphical plot the predicted values against the actual values. Regression curves are used to evaluate the performance of regression models.

These metrics can be used to test the performance of your model before final deployment.

Model deployment to LabVIEW

ANNHUB allows you to export your models in .zip format. The zip file contains all the necessary information, such as the model's architecture, weights, and biases, to deploy in LabVIEW using ANSDL

Export_TrainedNeuralNetworkModel_InANNHUB