Microsoft has announced a new version of its machine learning framework “ML.NET (v1.0)” and Model Builder for Visual Studio.
ML.NET is an open-source and cross-platform machine learning framework developed to provide model-based machine learning capabilities to .NET developers across the globe.
“Using ML.NET, developers can leverage their existing tools and skillsets to develop and infuse custom AI into their applications by creating custom machine learning models for common scenarios like Sentiment Analysis, Recommendation, Image Classification and more,” Microsoft writes in a blog post.
The ML.NET v1.1 brings a number of new features and enhancements, including a new algorithm based on Super-Resolution Deep Convolutional Network.
For ML.NET, Microsoft has improved the process of loading images in a model. Now, developers can load in-memory images and process them directly.
There is a new Anomaly Detection algorithm named ‘SrCnnAnomalyDetection’ which is based on a Super-Resolution Deep Convolutional Network. The main benefit of this algorithm is that it doesn’t require prior training.
A new component is added to Time Series NuGet package, that will be useful in predicting uncertain events that are impacted by different situations.
For Model Builder, Microsoft has added support for new scenarios as requested by customers. The updates include a new Issue Classification Template to categorize support issues, information accuracy in the Evaluate step, improvement in Code Generation step and feature that addresses multiple customer feedback.
The new issue classification template will allow users to classify tabular data into multiple classes. For instance, users can use the template to predict the issues on GitHub, route the customer support tickets, and classify emails into different categories.
Customers can get started with ML.NET here.