Reinforcement Learning in large action spaces is a challenging problem. Cooperative multi-agent reinforcement learning (MARL) exacerbates matters by imposing various constraints on communication and observability. In this work, we consider the fundamental hurdle affecting both value-based and policy-gradient approaches: an exponential blowup of the action space with the number of agents. For value-based methods, it poses challenges in accurately representing the optimal value function. For policy gradient methods, it makes training the critic difficult and exacerbates the problem of the lagging critic. We show that from a learning theory perspective, both problems can be addressed by accurately representing the associated action-value function with a low-complexity hypothesis class. This requires accurately modelling the agent interactions in a sample efficient way. To this end, we propose a novel tensorised formulation of the Bellman equation. This gives rise to our method Tesseract, which views the Q-function as a tensor whose modes correspond to the action spaces of different agents. Algorithms derived from Tesseract decompose the Q-tensor across agents and utilise low-rank tensor approximations to model agent interactions relevant to the task. We provide PAC analysis for Tesseract-based algorithms and highlight their relevance to the class of rich observation MDPs. Empirical results in different domains confirm Tesseract's gains in sample efficiency predicted by the theory.
In real-world multiagent systems, agents with different capabilities may join or leave without altering the team's overarching goals. Coordinating teams with such dynamic composition is challenging: the optimal team strategy varies with the composition. We propose COPA, a coach-player framework to tackle this problem. We assume the coach has a global view of the environment and coordinates the players, who only have partial views, by distributing individual strategies. Specifically, we 1) adopt the attention mechanism for both the coach and the players; 2) propose a variational objective to regularize learning; and 3) design an adaptive communication method to let the coach decide when to communicate with the players. We validate our methods on a resource collection task, a rescue game, and the StarCraft micromanagement tasks. We demonstrate zero-shot generalization to new team compositions. Our method achieves comparable or better performance than the setting where all players have a full view of the environment. Moreover, we see that the performance remains high even when the coach communicates as little as 13% of the time using the adaptive communication strategy.
Evolution in nature illustrates that the creatures' biological structure and their sensorimotor skills adapt to the environmental changes for survival. Likewise, the ability to morph and acquire new skills can facilitate an embodied agent to solve tasks of varying complexities. In this work, we introduce a data-driven approach where effective hand designs naturally emerge for the purpose of grasping diverse objects. Jointly optimizing morphology and control imposes computational challenges since it requires constant evaluation of a black-box function that measures the performance of a combination of embodiment and behavior. We develop a novel Bayesian Optimization algorithm that efficiently co-designs the morphology and grasping skills jointly through learned latent-space representations. We design the grasping tasks based on a taxonomy of three human grasp types: power grasp, pinch grasp, and lateral grasp. Through experimentation and comparative study, we demonstrate the effectiveness of our approach in discovering robust and cost-efficient hand morphologies for grasping novel objects.
Deep learning-based object pose estimators are often unreliable and overconfident especially when the input image is outside the training domain, for instance, with sim2real transfer. Efficient and robust uncertainty quantification (UQ) in pose estimators is critically needed in many robotic tasks. In this work, we propose a simple, efficient, and plug-and-play UQ method for 6-DoF object pose estimation. We ensemble 2-3 pre-trained models with different neural network architectures and/or training data sources, and compute their average pairwise disagreement against one another to obtain the uncertainty quantification. We propose four disagreement metrics, including a learned metric, and show that the average distance (ADD) is the best learning-free metric and it is only slightly worse than the learned metric, which requires labeled target data. Our method has several advantages compared to the prior art: 1) our method does not require any modification of the training process or the model inputs; and 2) it needs only one forward pass for each model. We evaluate the proposed UQ method on three tasks where our uncertainty quantification yields much stronger correlations with pose estimation errors than the baselines. Moreover, in a real robot grasping task, our method increases the grasping success rate from 35% to 90%.
We refine the OrbNet model to accurately predict energy, forces, and other response properties for molecules using a graph neural-network architecture based on features from low-cost approximated quantum operators in the symmetry-adapted atomic orbital basis. The model is end-to-end differentiable due to the derivation of analytic gradients for all electronic structure terms, and is shown to be transferable across chemical space due to the use of domain-specific features. The learning efficiency is improved by incorporating physically motivated constraints on the electronic structure through multi-task learning. The model outperforms existing methods on energy prediction tasks for the QM9 dataset and for molecular geometry optimizations on conformer datasets, at a computational cost that is thousand-fold or more reduced compared to conventional quantum-chemistry calculations (such as density functional theory) that offer similar accuracy.
We present a hierarchical framework that combines model-based control and reinforcement learning (RL) to synthesize robust controllers for a quadruped (the Unitree Laikago). The system consists of a high-level controller that learns to choose from a set of primitives in response to changes in the environment and a low-level controller that utilizes an established control method to robustly execute the primitives. Our framework learns a controller that can adapt to challenging environmental changes on the fly, including novel scenarios not seen during training. The learned controller is up to 85~percent more energy efficient and is more robust compared to baseline methods. We also deploy the controller on a physical robot without any randomization or adaptation scheme.
Humans have an inherent ability to learn novel concepts from only a few samples and generalize these concepts to different situations. Even though today's machine learning models excel with a plethora of training data on standard recognition tasks, a considerable gap exists between machine-level pattern recognition and human-level concept learning. To narrow this gap, the Bongard Problems (BPs) were introduced as an inspirational challenge for visual cognition in intelligent systems. Albeit new advances in representation learning and learning to learn, BPs remain a daunting challenge for modern AI. Inspired by the original one hundred BPs, we propose a new benchmark Bongard-LOGO for human-level concept learning and reasoning. We develop a program-guided generation technique to produce a large set of human-interpretable visual cognition problems in action-oriented LOGO language. Our benchmark captures three core properties of human cognition: 1) context-dependent perception, in which the same object may have disparate interpretations given different contexts; 2) analogy-making perception, in which some meaningful concepts are traded off for other meaningful concepts; and 3) perception with a few samples but infinite vocabulary. In experiments, we show that the state-of-the-art deep learning methods perform substantially worse than human subjects, implying that they fail to capture core human cognition properties. Finally, we discuss research directions towards a general architecture for visual reasoning to tackle this benchmark.