Revealing Vulnerabilities of Neural Networks in Parameter Learning and Defense Against Explanation-Aware Backdoors

arXiv:2403.16569v1 Announce Type: new
Abstract: Explainable Artificial Intelligence (XAI) strategies play a crucial part in increasing the understanding and trustworthiness of neural networks. Nonetheless, these techniques could potentially generate misleading explanations. Blinding attacks can drastically alter a machine learning algorithm's prediction and explanation, providing misleading information by adding visually unnoticeable artifacts into the input, while maintaining the model's accuracy. It poses a serious challenge in ensuring the reliability of XAI methods. To ensure the reliability of XAI methods poses a …

abstract algorithm artificial artificial intelligence arxiv attacks cs.cr cs.cv cs.lg defense explainable artificial intelligence generate information intelligence machine machine learning networks neural networks part prediction strategies type understanding vulnerabilities xai