Learning to Look: Seeking Information for Decision Making via Policy Factorization

Many robot manipulation tasks require active or interactive exploration behavior in order to be performed successfully. Such tasks are ubiquitous in embodied domains, where agents must actively search for the information necessary for each stage of a task, e.g., moving the head of the robot to find information relevant to manipulation, or in multi-robot domains, where one scout robot may search for the information that another robot needs to make informed decisions. We identify these tasks with a new type of problem, factorized Contextual Markov Decision Processes, and propose DISaM, a dual-policy solution composed of an information-seeking policy that explores the environment to find the relevant contextual information and an information-receiving policy that exploits the context to achieve the manipulation goal. This factorization allows us to train both policies separately, using the information-receiving one to provide reward to train the information-seeking policy. At test time, the dual agent balances exploration and exploitation based on the uncertainty the manipulation policy has on what the next best action is. We demonstrate the capabilities of our dual policy solution in five manipulation tasks that require information-seeking behaviors, both in simulation and in the real-world, where DISaM significantly outperforms existing methods. More information at https://robin-lab.cs.utexas.edu/learning2look/.

SkiLD: Unsupervised Skill Discovery Guided by Factor Interactions

Unsupervised skill discovery carries the promise that an intelligent agent can learn reusable skills through autonomous, reward-free environment interaction. Existing unsupervised skill discovery methods learn skills by encouraging distinguishable behaviors that cover diverse states. However, in complex environments with many state factors (e.g., household environments with many objects), learning skills that cover all possible states is impossible, and naively encouraging state diversity often leads to simple skills that are not ideal for solving downstream tasks. This work introduces Skill Discovery from Local Dependencies (Skild), which leverages state factorization as a natural inductive bias to guide the skill learning process. The key intuition guiding Skild is that skills that induce <b>diverse interactions</b> between state factors are often more valuable for solving downstream tasks. To this end, Skild develops a novel skill learning objective that explicitly encourages the mastering of skills that effectively induce different interactions within an environment. We evaluate Skild in several domains with challenging, long-horizon sparse reward tasks including a realistic simulated household robot domain, where Skild successfully learns skills with clear semantic meaning and shows superior performance compared to existing unsupervised reinforcement learning methods that only maximize state coverage.

Disentangled Unsupervised Skill Discovery for Efficient Hierarchical Reinforcement Learning

A hallmark of intelligent agents is the ability to learn reusable skills purely from unsupervised interaction with the environment. However, existing unsupervised skill discovery methods often learn entangled skills where one skill variable simultaneously influences many entities in the environment, making downstream skill chaining extremely challenging. We propose Disentangled Unsupervised Skill Discovery (DUSDi), a method for learning disentangled skills that can be efficiently reused to solve downstream tasks. DUSDi decomposes skills into disentangled components, where each skill component only affects one factor of the state space. Importantly, these skill components can be concurrently composed to generate low-level actions, and efficiently chained to tackle downstream tasks through hierarchical Reinforcement Learning. DUSDi defines a novel mutual-information-based objective to enforce disentanglement between the influences of different skill components, and utilizes value factorization to optimize this objective efficiently. Evaluated in a set of challenging environments, DUSDi successfully learns disentangled skills, and significantly outperforms previous skill discovery methods when it comes to applying the learned skills to solve downstream tasks. Code and skills visualization at jiahenghu.github.io/DUSDi-site/.

BUMBLE: Unifying Reasoning and Acting with Vision-Language Models for Building-wide Mobile Manipulation

To operate at a building scale, service robots must perform very long-horizon mobile manipulation tasks by navigating to different rooms, accessing different floors, and interacting with a wide and unseen range of everyday objects. We refer to these tasks as Building-wide Mobile Manipulation. To tackle these inherently long-horizon tasks, we introduce BUMBLE, a unified Vision-Language Model (VLM)-based framework integrating open-world RGBD perception, a wide spectrum of gross-to-fine motor skills, and dual-layered memory. Our extensive evaluation (90+ hours) indicates that BUMBLE outperforms multiple baselines in long-horizon building-wide tasks that require sequencing up to 12 ground truth skills spanning 15 minutes per trial. BUMBLE achieves 47.1% success rate averaged over 70 trials in different buildings, tasks, and scene layouts from different starting rooms and floors. Our user study demonstrates 22% higher satisfaction with our method than state-of-the-art mobile manipulation methods. Finally, we demonstrate the potential of using increasingly-capable foundation models to push performance further. For more information, see https://robin-lab.cs.utexas.edu/BUMBLE/

KinScene: Model-Based Mobile Manipulation of Articulated Scenes

Sequentially interacting with articulated objects is crucial for a mobile manipulator to operate effectively in everyday environments. To enable long-horizon tasks involving articulated objects, this study explores building scene-level articulation models for indoor scenes through autonomous exploration. While previous research has studied mobile manipulation with articulated objects by considering object kinematic constraints, it primarily focuses on individual-object scenarios and lacks extension to a scene-level context for task-level planning. To manipulate multiple object parts sequentially, the robot needs to reason about the resultant motion of each part and anticipate its impact on future actions.We introduce \ourtool{}, a full-stack approach for long-horizon manipulation tasks with articulated objects. The robot maps the scene, detects and physically interacts with articulated objects, collects observations, and infers the articulation properties. For sequential tasks, the robot plans a feasible series of object interactions based on the inferred articulation model. We demonstrate that our approach repeatably constructs accurate scene-level kinematic and geometric models, enabling long-horizon mobile manipulation in a real-world scene. Code and additional results are available at https://chengchunhsu.github.io/KinScene/

Beyond Shortsighted Navigation: Merging Best View Trajectory Planning with Robot Navigation

Gathering visual information effectively to monitor known environments is a key challenge in robotics. To be as efficient as human surveyors, robotic systems must continuously collect observational data required to complete their survey task. Inspection personnel instinctively know to look at relevant equipment that happens to be ``along the way.'' In this paper, we introduce a novel framework for continuous long-horizon viewpoint planning, for ground robots, applied to tasks involving patrolling, monitoring or visual data gathering in known environments. Our approach to Long Horizon Viewpoint Planning (LHVP), enables the robot to autonomously navigate and collect environmental data optimizing for coverage over the horizon of the patrol. Leveraging a quadruped's mobility and sensory capabilities, our LHVP framework plans patrol paths that account for coupling the viewpoint planner for the arm camera with the mobile base's navigation planner. The viewpath optimization algorithm seeks a balance between comprehensive environmental coverage and dynamically feasible movements, thus ensuring prolonged and effective operation in scenarios including monitoring, security surveillance, and disaster response. We validate our approach through simulations and in the real world and show that our LHVP significantly outperforms naive patrolling methods in terms of area coverage generating information-gathering trajectories for the robot arm. Our results indicate a promising direction for the deployment of mobile robots in long-term, autonomous surveying, and environmental data collection tasks, highlighting the potential of intelligent robotic systems in challenging real-world applications.

Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes

Reinforcement learning (RL), particularly its combination with deep neural networks referred to as deep RL (DRL), has shown tremendous promise across a wide range of applications, suggesting its potential for enabling the development of sophisticated robotic behaviors. Robotics problems, however, pose fundamental difficulties for the application of RL, stemming from the complexity and cost of interacting with the physical world. This article provides a modern survey of DRL for robotics, with a particular focus on evaluating the real-world successes achieved with DRL in realizing several key robotic competencies. Our analysis aims to identify the key factors underlying those exciting successes, reveal underexplored areas, and provide an overall characterization of the status of DRL in robotics. We highlight several important avenues for future work, emphasizing the need for stable and sample-efficient real-world RL paradigms, holistic approaches for discovering and integrating various competencies to tackle complex long-horizon, open-world tasks, and principled development and evaluation procedures. This survey is designed to offer insights for both RL practitioners and roboticists toward harnessing RL's power to create generally capable real-world robotic systems.

Natural Language Can Help Bridge the Sim2Real Gap

The main challenge in learning image-conditioned robotic policies is acquiring a visual representation conducive to low-level control. Due to the high dimensionality of the image space, learning a good visual representation requires a considerable amount of visual data. However, when learning in the real world, data is expensive. Sim2Real is a promising paradigm for overcoming data scarcity in the real-world target domain by using a simulator to collect large amounts of cheap data closely related to the target task. However, it is difficult to transfer an image-conditioned policy from sim to real when the domains are very visually dissimilar. To bridge the sim2real visual gap, we propose using natural language descriptions of images as a unifying signal across domains that captures the underlying task-relevant semantics. Our key insight is that if two image observations from different domains are labeled with similar language, the policy should predict similar action distributions for both images. We demonstrate that training the image encoder to predict the language description or the distance between descriptions of a sim or real image serves as a useful, data-efficient pretraining step that helps learn a domain-invariant image representation. We can then use this image encoder as the backbone of an IL policy trained simultaneously on a large amount of simulated and a handful of real demonstrations. Our approach outperforms widely used prior sim2real methods and strong vision-language pretraining baselines like CLIP and R3M by 25 to 40%.

BEHAVIOR Vision Suite: Customizable Dataset Generation via Simulation

The systematic evaluation and understanding of computer vision models under varying conditions require large amounts of data with comprehensive and customized labels, which real-world vision datasets rarely satisfy. While current synthetic data generators offer a promising alternative, particularly for embodied AI tasks, they often fall short for computer vision tasks due to low asset and rendering quality, limited diversity, and unrealistic physical properties. We introduce the BEHAVIOR Vision Suite (BVS), a set of tools and assets to generate fully customized synthetic data for systematic evaluation of computer vision models, based on the newly developed embodied AI benchmark, BEHAVIOR-1K. BVS supports a large number of adjustable parameters at the scene level (e.g., lighting, object placement), the object level (e.g., joint configuration, attributes such as "filled" and "folded"), and the camera level (e.g., field of view, focal length). Researchers can arbitrarily vary these parameters during data generation to perform controlled experiments. We showcase three example application scenarios: systematically evaluating the robustness of models across different continuous axes of domain shift, evaluating scene understanding models on the same set of images, and training and evaluating simulation-to-real transfer for a novel vision task: unary and binary state prediction. Project website: https://behavior-vision-suite.github.io/

ScrewMimic: Bimanual Imitation from Human Videos with Screw Space Projection

Bimanual manipulation is a longstanding challenge in robotics due to the large number of degrees of freedom and the strict spatial and temporal synchronization required to generate meaningful behavior. Humans learn bimanual manipulation skills by watching other humans and by refining their abilities through play. In this work, we aim to enable robots to learn bimanual manipulation behaviors from human video demonstrations and fine-tune them through interaction. Inspired by seminal work in psychology and biomechanics, we propose modeling the interaction between two hands as a serial kinematic linkage -- as a screw motion, in particular, that we use to define a new action space for bimanual manipulation: screw actions. We introduce ScrewMimic, a framework that leverages this novel action representation to facilitate learning from human demonstration and self-supervised policy fine-tuning. Our experiments demonstrate that ScrewMimic is able to learn several complex bimanual behaviors from a single human video demonstration, and that it outperforms baselines that interpret demonstrations and fine-tune directly in the original space of motion of both arms. For more information and video results, https://robin-lab.cs.utexas.edu/ScrewMimic/