In this paper, we present a novel formulation for training neural networks that considers the distance of data observations to the decision boundary.
Oct 13, 2020 · We present a novel formulation for training neural networks that considers the distance of data points to the decision boundary.
FaiR-N uses a novel distance to the boundary formulation in order to: - reduce the disparity in the average ability of recourse (i.e. the change needed to ...
We present a novel method for fair and robust training of neural networks on tabular data. We implement a novel distance metric and minimize the average ...
In this paper, we present a novel formulation for training neural networks that considers the distance of data observations to the decision boundary such that ...
Fairness in machine learning is crucial when individuals are subject to automated decisions made by models in high-stake domains.
In this paper, we develop an adaptive dynamic programming-based robust tracking control for a class of continuous-time matched uncertain nonlinear systems. By ...
Oct 12, 2020 · We demonstrate that training with this loss yields more fair and robust neural networks with similar accuracies to models trained without it.
FaiR-N: Fair and Robust Neural Networks for Structured Data. May 19, 2021. Speakers. SS · Shubham Sharma. Speaker · 0 followers. Follow. AG ...
People also ask
Are neural networks robust?
Which neural network tends to perform better with amount of data?
Which neural network is best for tabular data?
What is a neural network in AI?
▷ Most neural networks are vulnerable to adversarial examples. ▷ Most empirical methods to prevent them do not work very well. ▷ Lipschitz-networks trained with ...