Reducing the Deployment-Time Inference Control Costs of Deep Reinforcement Learning Agents via an Asymmetric Architecture
This work addresses energy limitations for mobile robots, but it is incremental as it builds on existing DRL methods with a cost-reduction focus.
The paper tackles the problem of high inference costs in deep reinforcement learning (DRL) for mobile robots by proposing an asymmetric architecture that switches between expensive and economic policies, reducing costs while maintaining performance in robotic control tasks.
Deep reinforcement learning (DRL) has been demonstrated to provide promising results in several challenging decision making and control tasks. However, the required inference costs of deep neural networks (DNNs) could prevent DRL from being applied to mobile robots which cannot afford high energy-consuming computations. To enable DRL methods to be affordable in such energy-limited platforms, we propose an asymmetric architecture that reduces the overall inference costs via switching between a computationally expensive policy and an economic one. The experimental results evaluated on a number of representative benchmark suites for robotic control tasks demonstrate that our method is able to reduce the inference costs while retaining the agent's overall performance.