Outlier Detection Using Generative Models with Theoretical Performance Guarantees
It addresses outlier detection in signal recovery for applications like imaging or data analysis, but appears incremental as it builds on existing generative model frameworks with new theoretical guarantees.
This paper tackles the problem of recovering signals from linear measurements contaminated with sparse outliers using generative models, establishing theoretical guarantees for reconstruction and showing that their approach outperforms traditional methods like Lasso and ℓ2 minimization in experiments.
This paper considers the problem of recovering signals modeled by generative models from linear measurements contaminated with sparse outliers. We propose an outlier detection approach for reconstructing the ground-truth signals modeled by generative models under sparse outliers. We establish theoretical recovery guarantees for reconstruction of signals using generative models in the presence of outliers, giving lower bounds on the number of correctable outliers. Our results are applicable to both linear generator neural networks and the nonlinear generator neural networks with an arbitrary number of layers. We propose an iterative alternating direction method of multipliers (ADMM) algorithm for solving the outlier detection problem via $\ell_1$ norm minimization, and a gradient descent algorithm for solving the outlier detection problem via squared $\ell_1$ norm minimization. We conduct extensive experiments using variational auto-encoder and deep convolutional generative adversarial networks, and the experimental results show that the signals can be successfully reconstructed under outliers using our approach. Our approach outperforms the traditional Lasso and $\ell_2$ minimization approach.