skip to main content


Title: Hierarchical Neural Dynamic Policies
We tackle the problem of generalization to unseen configurations for dynamic tasks in the real world while learning from high-dimensional image input. The family of nonlinear dynamical system-based methods have successfully demonstrated dynamic robot behaviors but have difficulty in generalizing to unseen configurations as well as learning from image inputs. Recent works approach this issue by using deep network policies and reparameterize actions to embed the structure of dynamical systems but still struggle in domains with diverse configurations of image goals, and hence, find it difficult to generalize. In this paper, we address this dichotomy by leveraging embedding the structure of dynamical systems in a hierarchical deep policy learning framework, called Hierarchical Neural Dynamical Policies (H-NDPs). Instead of fitting deep dynamical systems to diverse data directly, H-NDPs form a curriculum by learning local dynamical system-based policies on small regions in state-space and then distill them into a global dynamical system-based policy that operates only from high-dimensional images. H-NDPs additionally provide smooth trajectories, a strong safety benefit in the real world. We perform extensive experiments on dynamic tasks both in the real world (digit writing, scooping, and pouring) and simulation (catching, throwing, picking). We show that H-NDPs are easily integrated with both imitation as well as reinforcement learning setups and achieve state-of-the-art results.  more » « less
Award ID(s):
2024594
NSF-PAR ID:
10312794
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
Robotics: Science and Systems
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Learning a dynamical system requires stabilizing the unknown dynamics to avoid state blow-ups. However, the standard reinforcement learning (RL) methods lack formal stabilization guarantees, which limits their applicability for the control of real-world dynamical systems. We propose a novel policy optimization method that adopts Krasovskii's family of Lyapunov functions as a stability constraint. We show that solving this stability-constrained optimization problem using a primal-dual approach recovers a stabilizing policy for the underlying system even under modeling error. Combining this method with model learning, we propose a model-based RL framework with formal stability guarantees, Krasovskii-Constrained Reinforcement Learning (KCRL). We theoretically study KCRL with kernel-based feature representation in model learning and provide a sample complexity guarantee to learn a stabilizing controller for the underlying system. Further, we empirically demonstrate the effectiveness of KCRL in learning stabilizing policies in online voltage control of a distributed power system. We show that KCRL stabilizes the system under various real-world solar and electricity demand profiles, whereas standard RL methods often fail to stabilize. 
    more » « less
  2. null (Ed.)
    We investigate the ways in which a machine learning architecture known as Reservoir Computing learns concepts such as “similar” and “different” and other relationships between image pairs and generalizes these concepts to previously unseen classes of data. We present two Reservoir Computing architectures, which loosely resemble neural dynamics, and show that a Reservoir Computer (RC) trained to identify relationships between image pairs drawn from a subset of training classes generalizes the learned relationships to substantially different classes unseen during training. We demonstrate our results on the simple MNIST handwritten digit database as well as a database of depth maps of visual scenes in videos taken from a moving camera. We consider image pair relationships such as images from the same class; images from the same class with one image superposed with noise, rotated 90°, blurred, or scaled; images from different classes. We observe that the reservoir acts as a nonlinear filter projecting the input into a higher dimensional space in which the relationships are separable; i.e., the reservoir system state trajectories display different dynamical patterns that reflect the corresponding input pair relationships. Thus, as opposed to training in the entire high-dimensional reservoir space, the RC only needs to learns characteristic features of these dynamical patterns, allowing it to perform well with very few training examples compared with conventional machine learning feed-forward techniques such as deep learning. In generalization tasks, we observe that RCs perform significantly better than state-of-the-art, feed-forward, pair-based architectures such as convolutional and deep Siamese Neural Networks (SNNs). We also show that RCs can not only generalize relationships, but also generalize combinations of relationships, providing robust and effective image pair classification. Our work helps bridge the gap between explainable machine learning with small datasets and biologically inspired analogy-based learning, pointing to new directions in the investigation of learning processes. 
    more » « less
  3. null (Ed.)
    Robotic manipulation of deformable 1D objects such as ropes, cables, and hoses is challenging due to the lack of high-fidelity analytic models and large configuration spaces. Furthermore, learning end-to-end manipulation policies directly from images and physical interaction requires significant time on a robot and can fail to generalize across tasks. We address these challenges using interpretable deep visual representations for rope, extending recent work on dense object descriptors for robot manipulation. This facilitates the design of interpretable and transferable geometric policies built on top of the learned representations, decoupling visual reasoning and control. We present an approach that learns point-pair correspondences between initial and goal rope configurations, which implicitly encodes geometric structure, entirely in simulation from synthetic depth images. We demonstrate that the learned representation - dense depth object descriptors (DDODs) - can be used to manipulate a real rope into a variety of different arrangements either by learning from demonstrations or using interpretable geometric policies. In 50 trials of a knot-tying task with the ABB YuMi Robot, the system achieves a 66% knot-tying success rate from previously unseen configurations. See https://tinyurl.com/rope-learning for supplementary material and videos. 
    more » « less
  4. Reward signals in reinforcement learning are expensive to design and often require access to the true state which is not available in the real world. Common alternatives are usually demonstrations or goal images which can be labor-intensive to collect. On the other hand, text descriptions provide a general, natural, and low-effort way of communicating the desired task. However, prior works in learning text-conditioned policies still rely on rewards that are defined using either true state or labeled expert demonstrations. We use recent developments in building large-scale visuolanguage models like CLIP to devise a framework that generates the task reward signal just from goal text description and raw pixel observations which is then used to learn the task policy. We evaluate the proposed framework on control and robotic manipulation tasks. Finally, we distill the individual task policies into a single goal text conditioned policy that can generalize in a zero-shot manner to new tasks with unseen objects and unseen goal text descriptions. 
    more » « less
  5. Simulation parameter settings such as contact models and object geometry approximations are critical to training robust robotic policies capable of transferring from simulation to real-world deployment. Previous approaches typically handcraft distributions over such parameters (domain randomization), or identify parameters that best match the dynamics of the real environment (system identification). However, there is often an irreducible gap between simulation and reality: attempting to match the dynamics between simulation and reality across all states and tasks may be infeasible and may not lead to policies that perform well in reality for a specific task. Addressing this issue, we propose AdaptSim, a new task-driven adaptation framework for sim-to-real transfer that aims to optimize task performance in target (real) environments -- instead of matching dynamics between simulation and reality. First, we meta-learn an adaptation policy in simulation using reinforcement learning for adjusting the simulation parameter distribution based on the current policy's performance in a target environment. We then perform iterative real-world adaptation by inferring new simulation parameter distributions for policy training, using a small amount of real data. We perform experiments in three robotic tasks: (1) swing-up of linearized double pendulum, (2) dynamic table-top pushing of a bottle, and (3) dynamic scooping of food pieces with a spatula. Our extensive simulation and hardware experiments demonstrate AdaptSim achieving 1-3x asymptotic performance and ∼2x real data efficiency when adapting to different environments, compared to methods based on Sys-ID and directly training the task policy in target environments. 
    more » « less