Search Results for author:
Found 4 papers, 1 papers with code
Unveiling and Mitigating Backdoor Vulnerabilities based on Unlearning Weight Changes and Backdoor Activeness
Recently, without poisoned data, unlearning models with clean data and then learning a pruning mask have contributed to backdoor defense.
TAOTF: A Two-stage Approximately Orthogonal Training Framework in Deep Neural Networks
In the first stage, we propose a novel algorithm called polar decomposition-based orthogonal initialization (PDOI) to find a good initialization for the orthogonal optimization.
Data-free Backdoor Removal based on Channel Lipschitzness
Pruning these channels was then shown to be effective in mitigating the backdoor behaviors.
Point cloud denoising based on tensor Tucker decomposition
In this paper, we propose a new algorithm for point cloud denoising based on the tensor Tucker decomposition.
