Transform2Act: Learning a Transform-and-Control Policy for Efficient Agent Design
This addresses the problem of efficient agent design for robotics and simulation, offering a novel approach that improves sample efficiency and performance, though it is incremental in combining design and control optimization.
The paper tackles the challenge of finding optimal agent designs (skeletal structure and joint attributes) for given functions by incorporating design into decision-making, learning a conditional policy that modifies designs and then controls them, resulting in significantly outperforming prior methods in convergence speed and final performance, with automatic discovery of plausible designs like giraffes and spiders.
An agent's functionality is largely determined by its design, i.e., skeletal structure and joint attributes (e.g., length, size, strength). However, finding the optimal agent design for a given function is extremely challenging since the problem is inherently combinatorial and the design space is prohibitively large. Additionally, it can be costly to evaluate each candidate design which requires solving for its optimal controller. To tackle these problems, our key idea is to incorporate the design procedure of an agent into its decision-making process. Specifically, we learn a conditional policy that, in an episode, first applies a sequence of transform actions to modify an agent's skeletal structure and joint attributes, and then applies control actions under the new design. To handle a variable number of joints across designs, we use a graph-based policy where each graph node represents a joint and uses message passing with its neighbors to output joint-specific actions. Using policy gradient methods, our approach enables joint optimization of agent design and control as well as experience sharing across different designs, which improves sample efficiency substantially. Experiments show that our approach, Transform2Act, outperforms prior methods significantly in terms of convergence speed and final performance. Notably, Transform2Act can automatically discover plausible designs similar to giraffes, squids, and spiders. Code and videos are available at https://sites.google.com/view/transform2act.