ダウンロード数: 53
このアイテムのファイル:
ファイル | 記述 | サイズ | フォーマット | |
---|---|---|---|---|
ACCESS.2021.3121751.pdf | 2.1 MB | Adobe PDF | 見る/開く |
タイトル: | Deep Adversarial Reinforcement Learning With Noise Compensation by Autoencoder |
著者: | Ohashi, Kohei Nakanishi, Kosuke Sasaki, Wataru Yasui, Yuji Ishii, Shin |
著者名の別形: | 大橋, 康平 中西, 康輔 佐々木, 航 石井, 信 |
キーワード: | Deep reinforcement learning adversarial learning robustness regularization automatic vehicle control |
発行日: | 2021 |
出版者: | Institute of Electrical and Electronics Engineers (IEEE) |
誌名: | IEEE Access |
巻: | 9 |
開始ページ: | 143901 |
終了ページ: | 143912 |
抄録: | We present a new adversarial learning method for deep reinforcement learning (DRL). Based on this method, robust internal representation in a deep Q-network (DQN) was introduced by applying adversarial noise to disturb the DQN policy; however, it was compensated for by the autoencoder network. In particular, we proposed the use of a new type of adversarial noise: it encourages the policy to choose the worst action leading to the worst outcome at each state. When the proposed method, called deep Q-W-network regularized with an autoencoder (DQWAE), was applied to seven different games in an Atari 2600, the results were convincing. DQWAE exhibited greater robustness against the random/adversarial noise added to the input and accelerated the learning process more than the baseline DQN. When applied to a realistic automatic driving simulation, the proposed DRL method was found to be effective at rendering the acquired policy robust against random/adversarial noise. |
著作権等: | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. |
URI: | http://hdl.handle.net/2433/277586 |
DOI(出版社版): | 10.1109/ACCESS.2021.3121751 |
出現コレクション: | 学術雑誌掲載論文等 |
このアイテムは次のライセンスが設定されています: クリエイティブ・コモンズ・ライセンス