Visualize an Object Classification Network with Crow for Grasshopper
Monday 19, February 2018
Crow exposes artificial neural networks in Grasshopper, offering supervisde learning through backpropagation networks and unsupervised learning through self-organizing maps. As mentioned by Pennjamin (developer of FlexHopper and Crow), "The new feature let's you visualise network training in real time. [In hopes of] better intuition on how to setup a good network topology." The video shows the classification of fruit-like shapes, but more complex tasks could be achieved with this tool.
It's exciting to start seeing the introduction of machine intelligence algorithms in tools like Grasshopper and Dynamo.