Dropout Feature Ranking for Deep Learning Models
By Chun-Hao Kingsley Chang, Ladislav Rampášek, Anna Goldenberg. 2017
53 citations recorded by the source corpus.
Deep neural networks (DNNs) achieve state-of-the-art results in a variety of domains. Unfortunately, DNNs are notorious for their non-interpretability, and thus limit their applicability in hypothesis-driven domains such as biology and healthcare. Moreover, in the resource-constraint setting, it is critical to design tests relying on fewer more informative features leading to high accuracy performance within reasonable budget. We aim to close this gap by proposing a new general feature ranking method for deep learning. We show that our simple yet effective method performs on par or compares favorably to eight strawman, classical and deep-learning feature ranking methods in two simulations and five very different datasets on tasks ranging from classification to regression, in both static and time series scenarios. We also illustrate the use of our method on a drug response dataset and show that it identifies genes relevant to the drug-response.
Subfields: Artificial Intelligence, Molecular Biology
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Congzheng Song
- Vitaly Shmatikov
- Emiliano De Cristofaro
- Luca Melis
- Reza Shokri
- Yang Zhang
- Prateek Mittal
- Amir Houmansadr
In Molecular Biology
Citations in the corpus
- Exploiting Unintended Feature Leakage in Collaborative Learning (2018, background): “Dropout stochastically removes features at every collaborative training step, thus yielding more informative features (similar to feature bagging [7, 26]) and increasing variance between participants’ updates.”
- Inference Attacks Against Collaborative Learning (2018, background): “ slightly decreasing the accuracy of the joint model. Dropout stochastically removes features at every collaborative training step, thus yielding more informative features (similar to feature bagging [7,26]) and increasing variance between participants’ updates. 8.4Participant-level differential pri”