An official website of the United States government
Reward functions are a common way to specify the objective of a robot. As designing reward functions can be extremely challenging, a more promising approach is to directly learn reward functions from human teachers. Importantly, data from human teachers can be collected either passively or actively in a variety of forms: passive data sources include demonstrations (e.g., kinesthetic guidance), whereas preferences (e.g., comparative rankings) are actively elicited. Prior research has independently applied reward learning to these different data sources. However, there exist many domains where multiple sources are complementary and expressive. Motivated by this general problem, we present a framework to integrate multiple sources of information, which are either passively or actively collected from human users. In particular, we present an algorithm that first utilizes user demonstrations to initialize a belief about the reward function, and then actively probes the user with preference queries to zero-in on their true reward. This algorithm not only enables us combine multiple data sources, but it also informs the robot when it should leverage each type of information. Further, our approach accounts for the human’s ability to provide data: yielding user-friendly preference queries which are also theoretically optimal. Our extensive simulated experiments and user studies on a Fetch mobile manipulator demonstrate the superiority and the usability of our integrated framework.
more »
« less
Unified Learning from Demonstrations, Corrections, and Preferences during Physical Human–Robot Interaction
Humans can leverage physical interaction to teach robot arms. This physical interaction takes multiple forms depending on the task, the user, and what the robot has learned so far. State-of-the-art approaches focus on learning from a single modality, or combine some interaction types. Some methods do so by assuming that the robot has prior information about the features of the task and the reward structure. By contrast, in this article, we introduce an algorithmic formalism that unites learning from demonstrations, corrections, and preferences. Our approach makes no assumptions about the tasks the human wants to teach the robot; instead, we learn a reward model from scratch by comparing the human’s input to nearby alternatives, i.e., trajectories close to the human’s feedback. We first derive a loss function that trains an ensemble of reward models to match the human’s demonstrations, corrections, and preferences. The type and order of feedback is up to the human teacher: We enable the robot to collect this feedback passively or actively. We then apply constrained optimization to convert our learned reward into a desired robot trajectory. Through simulations and a user study, we demonstrate that our proposed approach more accurately learns manipulation tasks from physical human interaction than existing baselines, particularly when the robot is faced with new or unexpected objectives. Videos of our user study are available at https://youtu.be/FSUJsTYvEKU
more »
« less
- Award ID(s):
- 2129201
- PAR ID:
- 10567707
- Publisher / Repository:
- ACM Transactions on Human-Robot Interaction
- Date Published:
- Journal Name:
- ACM Transactions on Human-Robot Interaction
- Volume:
- 13
- Issue:
- 3
- ISSN:
- 2573-9522
- Page Range / eLocation ID:
- 1 to 25
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Learning from Demonstration (LfD) approaches empower end-users to teach robots novel tasks via demonstrations of the desired behaviors, democratizing access to robotics. However, current LfD frameworks are not capable of fast adaptation to heterogeneous human demonstrations nor the large-scale deployment in ubiquitous robotics applications. In this paper, we propose a novel LfD framework, Fast Lifelong Adaptive Inverse Reinforcement learning (FLAIR). Our approach (1) leverages learned strategies to construct policy mixtures for fast adaptation to new demonstrations, allowing for quick end-user personalization, (2) distills common knowledge across demonstrations, achieving accurate task inference; and (3) expands its model only when needed in lifelong deployments, maintaining a concise set of prototypical strategies that can approximate all behaviors via policy mixtures. We empirically validate that FLAIR achieves adaptability (i.e., the robot adapts to heterogeneous, user-specific task preferences), efficiency (i.e., the robot achieves sample-efficient adaptation), and scalability (i.e., the model grows sublinearly with the number of demonstrations while maintaining high performance). FLAIR surpasses benchmarks across three control tasks with an average 57% improvement in policy returns and an average 78% fewer episodes required for demonstration modeling using policy mixtures. Finally, we demonstrate the success of FLAIR in a table tennis task and find users rate FLAIR as having higher task (p < .05) and personalization (p < .05) performance.more » « less
-
Assistive robot arms can help humans by partially automating their desired tasks. Consider an adult with motor impairments controlling an assistive robot arm to eat dinner. The robot can reduce the number of human inputs — and how precise those inputs need to be — by recognizing what the human wants (e.g., a fork) and assisting for that task (e.g., moving towards the fork). Prior research has largely focused on learning the human’s task and providing meaningful assistance. But as the robot learns and assists, we also need to ensure that the human understands the robot’s intent (e.g., does the human know the robot is reaching for a fork?). In this paper, we study the effects of communicating learned assistance from the robot back to the human operator. We do not focus on the specific interfaces used for communication. Instead, we develop experimental and theoretical models of a) how communication changes the way humans interact with assistive robot arms, and b) how robots can harness these changes to better align with the human’s intent. We first conduct online and in-person user studies where participants operate robots that provide partial assistance, and we measure how the human’s inputs change with and without communication. With communication, we find that humans are more likely to intervene when the robot incorrectly predicts their intent, and more likely to release control when the robot correctly understands their task. We then use these findings to modify an established robot learning algorithm so that the robot can correctly interpret the human’s inputs when communication is present. Our results from a second in-person user study suggest that this combination of communication and learning outperforms assistive systems that isolate either learning or communication. See videos here: https://youtu.be/BET9yuVTVU4more » « less
-
We study the problem of cross-embodiment inverse reinforcement learning, where we wish to learn a reward function from video demonstrations in one or more embodiments and then transfer the learned reward to a different embodiment (e.g., different action space, dynamics, size, shape, etc.). Learning reward functions that transfer across embodiments is important in settings such as teaching a robot a policy via human video demonstrations or teaching a robot to imitate a policy from another robot with a different embodiment. However, prior work has only focused on cases where near-optimal demonstrations are available, which is often difficult to ensure. By contrast, we study the setting of cross-embodiment reward learning from mixed-quality demonstrations. We demonstrate that prior work struggles to learn generalizable reward representations when learning from mixed-quality data. We then analyze several techniques that leverage human feedback for representation learning and alignment to enable effective cross-embodiment learning. Our results give insight into how different representation learning techniques lead to qualitatively different reward shaping behaviors and the importance of human feedback when learning from mixed-quality, mixed-embodiment data.more » « less
-
Reward learning as a method for inferring human intent and preferences has been studied extensively. Prior approaches make an implicit assumption that the human maintains a correct belief about the robot's domain dynamics. However, this may not always hold since the human's belief may be biased, which can ultimately lead to a misguided estimation of the human's intent and preferences, which is often derived from human feedback on the robot's behaviors. In this paper, we remove this restrictive assumption by considering that the human may have an inaccurate understanding of the robot. We propose a method called Generalized Reward Learning with biased beliefs about domain dynamics (GeReL) to infer both the reward function and human's belief about the robot in a Bayesian setting based on human ratings. Due to the complex forms of the posteriors, we formulate it as a variational inference problem to infer the posteriors of the parameters that govern the reward function and human's belief about the robot simultaneously. We evaluate our method in a simulated domain and with a user study where the user has a bias based on the robot's appearances. The results show that our method can recover the true human preferences while subject to such biased beliefs, in contrast to prior approaches that could have misinterpreted them completely.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1145/3623384
