Owl is a machine-learning oriented library for data processing. It’s build from scratch in .NET, greatly reduces the daily hurdles of data preparation, clean-up and visualization. The data is stored internally as arrays, exposing the raw data if necessary. This approach removes any slow-down caused by the nature of the Interfaces. Currently it is equipped with Extensions enabling it to work directly with the Accord framework (formely AForge).
An example of how it can help with visualization is provided in the picture below. The initially random weights are changed over time which is directly visualized as a colorized surface – the surface itself represents the network function.