LassoNet: Neural Networks with Feature Sparsity

Ismael Lemhadri Feng Ruan Louis Abraham Rob Tibshirani

LassoNet is a method for feature selection in neural networks, to enhance interpretability of the final network.


LassoNet in 2 minutes


Installation

pip install lassonet

Examples


Tips

LassoNet sometimes require fine tuning. For optimal performance, consider:

Intro video

Talk



Citation

The algorithms and method used in this package came primarily out of research in Rob Tibshirani's lab at Stanford University. If you use LassoNet in your research we would appreciate a citation to the paper:
    @article{JMLR:v22:20-848,
      author  = {Ismael Lemhadri and Feng Ruan and Louis Abraham and Robert Tibshirani},
      title   = {LassoNet: A Neural Network with Feature Sparsity},
      journal = {Journal of Machine Learning Research},
      year    = {2021},
      volume  = {22},
      number  = {127},
      pages   = {1-29},
      url     = {http://jmlr.org/papers/v22/20-848.html}
    }