We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Improving Molecular Graph Neural Network Explainability with Orthonormalization and Induced Sparsity

Abstract: Rationalizing which parts of a molecule drive the predictions of a molecular graph convolutional neural network (GCNN) can be difficult. To help, we propose two simple regularization techniques to apply during the training of GCNNs: Batch Representation Orthonormalization (BRO) and Gini regularization. BRO, inspired by molecular orbital theory, encourages graph convolution operations to generate orthonormal node embeddings. Gini regularization is applied to the weights of the output layer and constrains the number of dimensions the model can use to make predictions. We show that Gini and BRO regularization can improve the accuracy of state-of-the-art GCNN attribution methods on artificial benchmark datasets. In a real-world setting, we demonstrate that medicinal chemists significantly prefer explanations extracted from regularized models. While we only study these regularizers in the context of GCNNs, both can be applied to other types of neural networks
Comments: Accepted to ICML 2021
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2105.04854 [stat.ML]
  (or arXiv:2105.04854v2 [stat.ML] for this version)

Submission history

From: Ryan Henderson [view email]
[v1] Tue, 11 May 2021 08:13:34 GMT (2393kb,D)
[v2] Mon, 31 May 2021 12:07:30 GMT (2395kb,D)

Link back to: arXiv, form interface, contact.