Pytorch implementation of Neural Network with Dynamically Evolved Capacity (NADINE) published in CIKM 2019 titled " Automatic Construction of Multi-layer Perceptron Network from Streaming Examples. CIKM 2019: 1171-1180"
datasets can be downloaded from "https://drive.google.com/drive/folders/1Kdkcdagctx1tX3bGIPo85XWIYKLAoMww?usp=sharing"
To run the code, one can go directly to .ipynb files. For example, if one wants to try NADINE in SEA problem, one can go to NADINE_classification_sea.ipynb.
Our numerical results can be seen in the .html files.