A3C-S: Automated Agent Accelerator Co-Search towards Efficient Deep Reinforcement Learning

2021/06/11 by Yonggan Fu, Fu, Yonggan, Yongan Zhang +7
Computer Science · Engineering · #Evolutionary Algorithms and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Modular Robots and Swarm Intelligence #Reinforcement Learning in Robotics

paper · pdf · doi:10.48550/arxiv.2106.06577

openalex publication_date 2021/06/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Driven by the explosive interest in applying deep reinforcement learning (DRL) agents to numerous real-time control and decision-making applications, there has been a growing demand to deploy DRL agents to empower daily-life intelligent devices, while the prohibitive complexity of DRL stands at odds with limited on-device resources. In this work, we propose an Automated Agent Accelerator Co-Search (A3C-S) framework, which to our best knowledge is the first to automatically co-search the optimally matched DRL agents and accelerators that maximize both test scores and hardware efficiency. Extensive experiments consistently validate the superiority of our A3C-S over state-of-the-art techniques.

Citations

• Categorical Reparameterization with Gumbel-Softmax
• DARTS: Differentiable Architecture Search
• ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
• AutoGAN-Distiller: Searching to Compress Generative Adversarial Networks
• Deep Reinforcement Learning: A Brief Survey

Related

• Auto-Agent-Distiller: Towards Efficient Deep Reinforcement Learning Agents via Neural Architecture Search
• Multi-Agent Reinforcement Learning for Sample-Efficient Deep Neural Network Mapping
• Optimizing the Neural Architecture of Reinforcement Learning Agents
• MARCO: Hardware-Aware Neural Architecture Search for Edge Devices with Multi-Agent Reinforcement Learning and Conformal Prediction Filtering
• Distributed Deep Reinforcement Learning: Learn how to play Atari games\n in 21 minutes