Gradient Band-based Adversarial Training for Generalized Attack Immunity of A3C Path Finding

Authors: Tong Chen, Wenjia Niu, Yingxiao Xiang, Xiaoxuan Bai, Jiqiang Liu, Zhen Han, Gang Li
Published: 2018-07-18

Source: https://arxiv.org/abs/1807.06752

PDF: https://arxiv.org/pdf/1807.06752

AIにより推定されたラベル

ポイズニング 敵対的攻撃 敵対的学習

※ こちらのラベルはAIによって自動的に追加されました。そのため、正確でないことがあります。
詳細は文献データベースについてをご覧ください。

Abstract

As adversarial attacks pose a serious threat to the security of AI system in practice, such attacks have been extensively studied in the context of computer vision applications. However, few attentions have been paid to the adversarial research on automatic path finding. In this paper, we show dominant adversarial examples are effective when targeting A3C path finding, and design a Common Dominant Adversarial Examples Generation Method (CDG) to generate dominant adversarial examples against any given map. In addition, we propose Gradient Band-based Adversarial Training, which trained with a single randomly choose dominant adversarial example without taking any modification, to realize the “1:N” attack immunity for generalized dominant adversarial examples. Extensive experimental results show that, the lowest generation precision for CDG algorithm is 91.91 Adversarial Training is 93.89 generalized attack immunity of A3C path finding with a high confidence.