Salta al contenuto
logo-cnr logo-cnr

Dati Aperti bibLIoteca Area bologna

PUBLICATIONS

Modifiche

View changes from to

On 22 marzo 2025 23:00:12 UTC, Gravatar Marco Lamorte:

  • Updated description of resource 2024_MARKO RUMAN in PUBLICATIONS from

    **Title:** Knowledge Transfer in Deep Reinforcement Learning via an RL-Specific GAN-Based Correspondence Function. **Author(s):** MARKO RUMAN, TATIANA V. GUY. **DOI:** 10.1109/ACCESS.2024.3497589. **Publication Date:** 13 November 2024. **Resource Type:** Research Paper. **Format:** PDF. **Working Group:** WG1-WG2. **Affiliation(s):** A) Department of Adaptive Systems, Institute of Information Theory and Automation; B) Czech Academy of Sciences, 182 00 Prague, Czech Republic; C) Department of Information Engineering, Faculty of Economics and Management, Czech University of Life Sciences, 165 00 Prague, Czech Republic. **Open Access Status:** Open. **Keywords:** **Description/Abstract:** Deep reinforcement learning has demonstrated superhuman performance in complex decisionmaking tasks, but it struggles with generalization and knowledge reuse—key aspects of true intelligence. This article introduces a novel approach that modifies Cycle Generative Adversarial Networks specifically for reinforcement learning, enabling effective one-to-one knowledge transfer between two tasks. Our method enhances the loss function with two new components: model loss, which captures dynamic relationships between source and target tasks, and Q-loss, which identifies states significantly influencing the target decision policy. Tested on the 2-D Atari game Pong, our method achieved 100% knowledge transfer in identical tasks and either 100% knowledge transfer or a 30% reduction in training time for a rotated task, depending on the network architecture. In contrast, using standard Generative Adversarial Networks or Cycle Generative Adversarial Networks led to worse performance than training from scratch in the majority of cases. The results demonstrate that the proposed method ensured enhanced knowledge generalization in deep reinforcement learning.
    to
    **Title:** Knowledge Transfer in Deep Reinforcement Learning via an RL-Specific GAN-Based Correspondence Function. **Author(s):** MARKO RUMAN, TATIANA V. GUY. **DOI:** 10.1109/ACCESS.2024.3497589. **Publication Date:** 13 November 2024. **Resource Type:** Research Paper. **Format:** PDF. **Working Group:** WG1-WG2. **Affiliation(s):** A) Department of Adaptive Systems, Institute of Information Theory and Automation; B) Czech Academy of Sciences, 182 00 Prague, Czech Republic; C) Department of Information Engineering, Faculty of Economics and Management, Czech University of Life Sciences, 165 00 Prague, Czech Republic. **Access Status:** Open. **Keywords:** **Description:** Deep reinforcement learning has demonstrated superhuman performance in complex decisionmaking tasks, but it struggles with generalization and knowledge reuse—key aspects of true intelligence. This article introduces a novel approach that modifies Cycle Generative Adversarial Networks specifically for reinforcement learning, enabling effective one-to-one knowledge transfer between two tasks. Our method enhances the loss function with two new components: model loss, which captures dynamic relationships between source and target tasks, and Q-loss, which identifies states significantly influencing the target decision policy. Tested on the 2-D Atari game Pong, our method achieved 100% knowledge transfer in identical tasks and either 100% knowledge transfer or a 30% reduction in training time for a rotated task, depending on the network architecture. In contrast, using standard Generative Adversarial Networks or Cycle Generative Adversarial Networks led to worse performance than training from scratch in the majority of cases. The results demonstrate that the proposed method ensured enhanced knowledge generalization in deep reinforcement learning.