Deep reinforcement learning (RL) has led to encouraging successes in many challenging control tasks. However, a deep RL model lacks interpretability due to the difficulty of identifying how the model's control logic relates to its network structure. Programmatic policies structured in more interpretable representations emerge as a promising solution. Yet two shortcomings remain: First, synthesizing programmatic policies requires optimizing over the discrete and non-differentiable search space of program architectures. Previous works are suboptimal because they only enumerate program architectures greedily guided by a pretrained RL oracle. Second, these works do not exploit compositionality, an important programming concept, to reuse and compose primitive functions to form a complex function for new tasks. Our first contribution is a programmatically interpretable RL framework that conducts program architecture search on top of a continuous relaxation of the architecture space defined by programming language grammar rules. Our algorithm allows policy architectures to be learned with policy parameters via bilevel optimization using efficient policy-gradient methods, and thus does not require a pretrained oracle. Our second contribution is improving programmatic policies to support compositionality by integrating primitive functions learned to grasp task-agnostic skills as a composite program to solve novel RL problems. Experiment results demonstrate that our algorithm excels in discovering optimal programmatic policies that are highly interpretable.
more »
« less
This content will become publicly available on June 20, 2025
Reward-Guided Synthesis of Intelligent Agents with Control Structures
Deep reinforcement learning (RL) has led to encouraging successes in numerous challenging robotics applications. However, the lack of inductive biases to support logic deduction and generalization in the representation of a deep RL model causes it less effective in exploring complex long-horizon robot-control tasks with sparse reward signals. Existing program synthesis algorithms for RL problems inherit the same limitation, as they either adapt conventional RL algorithms to guide program search or synthesize robot-control programs to imitate an RL model. We propose ReGuS, a reward-guided synthesis paradigm, to unlock the potential of program synthesis to overcome the exploration challenges. We develop a novel hierarchical synthesis algorithm with decomposed search space for loops, on-demand synthesis of conditional statements, and curriculum synthesis for procedure calls, to effectively compress the exploration space for long-horizon, multi-stage, and procedural robot-control tasks that are difficult to address by conventional RL techniques. Experiment results demonstrate that ReGuS significantly outperforms state-of-the-art RL algorithms and standard program synthesis baselines on challenging robot tasks including autonomous driving, locomotion control, and object manipulation.
more » « less- Award ID(s):
- 2124155
- PAR ID:
- 10544039
- Publisher / Repository:
- ACM
- Date Published:
- Journal Name:
- Proceedings of the ACM on Programming Languages
- Volume:
- 8
- Issue:
- PLDI
- ISSN:
- 2475-1421
- Page Range / eLocation ID:
- 1730 to 1754
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Despite the potential of reinforcement learning (RL) for building general-purpose robotic systems, training RL agents to solve robotics tasks still remains challenging due to the difficulty of exploration in purely continuous action spaces. Addressing this problem is an active area of research with the majority of focus on improving RL methods via better optimization or more efficient exploration. An alternate but important component to consider improving is the interface of the RL algorithm with the robot. In this work, we manually specify a library of robot action primitives (RAPS), parameterized with arguments that are learned by an RL policy. These parameterized primitives are expressive, simple to implement, enable efficient exploration and can be transferred across robots, tasks and environments. We perform a thorough empirical study across challenging tasks in three distinct domains with image input and a sparse terminal reward. We find that our simple change to the action interface substantially improves both the learning efficiency and task performance irrespective of the underlying RL algorithm, significantly outperforming prior methods which learn skills from offline expert data. Code and videos at https://mihdalal.github.io/raps/more » « less
-
null (Ed.)A promising approach to solving challenging long-horizon tasks has been to extract behavior priors (skills) by fitting generative models to large offline datasets of demonstrations. However, such generative models inherit the biases of the underlying data and result in poor and unusable skills when trained on imperfect demonstration data. To better align skill extraction with human intent we present Skill Preferences (SkiP), an algorithm that learns a model over human preferences and uses it to extract human-aligned skills from offline data. After extracting human-preferred skills, SkiP also utilizes human feedback to solve downstream tasks with RL. We show that SkiP enables a simulated kitchen robot to solve complex multi-step manipulation tasks and substantially outperforms prior leading RL algorithms with human preferences as well as leading skill extraction algorithms without human preferences.more » « less
-
We propose a novel parameterized skill-learning algorithm that aims to learn transferable parameterized skills and synthesize them into a new action space that supports efficient learning in long-horizon tasks. We propose to leverage off-policy Meta-RL combined with a trajectory-centric smoothness term to learn a set of parameterized skills. Our agent can use these learned skills to construct a three-level hierarchical framework that models a Temporally-extended Parameterized Action Markov Decision Process. We empirically demonstrate that the proposed algorithms enable an agent to solve a set of difficult long-horizon (obstacle-course and robot manipulation) tasks.more » « less
-
more » « less
We present a strategy for simulation-to-real transfer, which builds on recent advances in robot skill decomposition. Rather than focusing on minimizing the simulation–reality gap, we propose a method for increasing the sample efficiency and robustness of existing simulation-to-real approaches which exploits hierarchy and online adaptation. Instead of learning a unique policy for each desired robotic task, we learn a diverse set of skills and their variations, and embed those skill variations in a continuously parameterized space. We then interpolate, search, and plan in this space to find a transferable policy which solves more complex, high-level tasks by combining low-level skills and their variations. In this work, we first characterize the behavior of this learned skill space, by experimenting with several techniques for composing pre-learned latent skills. We then discuss an algorithm which allows our method to perform long-horizon tasks never seen in simulation, by intelligently sequencing short-horizon latent skills. Our algorithm adapts to unseen tasks online by repeatedly choosing new skills from the latent space, using live sensor data and simulation to predict which latent skill will perform best next in the real world. Importantly, our method learns to control a real robot in joint-space to achieve these high-level tasks with little or no on-robot time, despite the fact that the low-level policies may not be perfectly transferable from simulation to real, and that the low-level skills were not trained on any examples of high-level tasks. In addition to our results indicating a lower sample complexity for families of tasks, we believe that our method provides a promising template for combining learning-based methods with proven classical robotics algorithms such as model-predictive control.
