Towards Practical Multi-Object Manipulation using Relational Reinforcement Learning
By R. Li, Allan Jabri, Trevor Darrell, Pulkit Agrawal. 2019
124 citations recorded by the source corpus.
Learning robotic manipulation tasks using reinforcement learning with sparse rewards is currently impractical due to the outrageous data requirements. Many practical tasks require manipulation of multiple objects, and the complexity of such tasks increases with the number of objects. Learning from a curriculum of increasingly complex tasks appears to be a natural solution, but unfortunately, does not work for many scenarios. We hypothesize that the inability of the state- of-the-art algorithms to effectively utilize a task curriculum stems from the absence of inductive biases for transferring knowledge from simpler to complex tasks. We show that graph-based relational architectures overcome this limitation and enable learning of complex tasks when provided with a simple curriculum of tasks with increasing numbers of objects. We demonstrate the utility of our framework on a simulated block stacking task. Starting from scratch, our agent learns to stack six blocks into a tower. Despite using step-wise sparse rewards, our method is orders of magnitude more data- efficient and outperforms the existing state-of-the-art method that utilizes human demonstrations. Furthermore, the learned policy exhibits zero-shot generalization, successfully stacking blocks into taller towers and previously unseen configurations such as pyramids, without any further training.
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
Citations in the corpus
- Discovering Diverse Multi-Agent Strategic Behavior via Reward Randomization (2021, background): “, 2019), φ is typically related to the bonus points for destroying each type of units; in robotics manipulation (Levine et al., 2016; Li et al., 2020; Yu et al., 2019), φ is often about the distance between the robot/object and its target position; in general multi-agent games (Lowe et al.”
- Rapid locomotion via reinforcement learning (2022, background): “Task difficulty is a function of both the system dynamics and the optimization algorithm, making manual curriculum design tedious and problem-dependent.”
- Deep Visual Reasoning: Learning to Predict Action Sequences for Task and Motion Planning from an Initial Scene Image (2020, methodology): “In [49] convolutional (CNN) and graph neural networks are utilized to learn a state representation for RL, similarly in [30].”
- Grounding Language to Autonomously-Acquired Skills via Goal Generation (2021, unknown): “, 2019) and a specific curiosity mechanism (Li et al., 2019); 2) they facilitate the grounding of language into skills and 3) they enable decoupling skill learning from language grounding, as observed in infants (Piaget, 1977).”