arXiv subjects

Jeong-Hoon Lee

Publications and source records attributed to Jeong-Hoon Lee.

3 recordsLinked to original sources

arXiv2023

Equivariant Descriptor Fields: SE(3)-Equivariant Energy-Based Models for End-to-End Visual Robotic Manipulation Learning

End-to-end learning for visual robotic manipulation is known to suffer from sample inefficiency, requiring large numbers of demonstrations. The spatial roto-translation equivariance, or the SE(3)-equivariance can be exploited to improve the sample efficiency for learning robotic manipulation. In this paper, we present SE(3)-equivariant models for visual robotic manipulation from point clouds that can be trained fully end-to-end. By utilizing the representation theory of the Lie group, we construct novel SE(3)-equivariant energy-based models that allow highly sample efficient end-to-end learning. We show that our models can learn from scratch without prior knowledge and yet are highly sample efficient (5~10 demonstrations are enough). Furthermore, we show that our models can generalize to tasks with (i) previously unseen target object poses, (ii) previously unseen target object instances of the category, and (iii) previously unseen visual distractors. We experiment with 6-DoF robotic manipulation tasks to validate our models' sample efficiency and generalizability. Codes are available at: https://github.com/tomato1mule/edf

cs.RO↗

arXiv2022

Hierarchical Primitive Composition: Simultaneous Activation of Skills with Inconsistent Action Dimensions in Multiple Hierarchies

Deep reinforcement learning has shown its effectiveness in various applications, providing a promising direction for solving tasks with high complexity. However, naively applying classical RL for learning a complex long-horizon task with a single control policy is inefficient. Thus, policy modularization tackles this problem by learning a set of modules that are mapped to primitives and properly orchestrating them. In this study, we further expand the discussion by incorporating simultaneous activation of the skills and structuring them into multiple hierarchies in a recursive fashion. Moreover, we sought to devise an algorithm that can properly orchestrate the skills with different action spaces via multiplicative Gaussian distributions, which highly increases the reusability. By exploiting the modularity, interpretability can also be achieved by observing the modules that are used in the new task if each of the skills is known. We demonstrate how the proposed scheme can be employed in practice by solving a pick and place task with a 6 DoF manipulator, and examine the effects of each property from ablation studies.

cs.LG↗

arXiv2009

Progressive Processing of Continuous Range Queries in Hierarchical Wireless Sensor Networks

In this paper, we study the problem of processing continuous range queries in a hierarchical wireless sensor network. Contrasted with the traditional approach of building networks in a "flat" structure using sensor devices of the same capability, the hierarchical approach deploys devices of higher capability in a higher tier, i.e., a tier closer to the server. While query processing in flat sensor networks has been widely studied, the study on query processing in hierarchical sensor networks has been inadequate. In wireless sensor networks, the main costs that should be considered are the energy for sending data and the storage for storing queries. There is a trade-off between these two costs. Based on this, we first propose a progressive processing method that effectively processes a large number of continuous range queries in hierarchical sensor networks. The proposed method uses the query merging technique proposed by Xiang et al. as the basis and additionally considers the trade-off between the two costs. More specifically, it works toward reducing the storage cost at lower-tier nodes by merging more queries, and toward reducing the energy cost at higher-tier nodes by merging fewer queries (thereby reducing "false alarms"). We then present how to build a hierarchical sensor network that is optimal with respect to the weighted sum of the two costs. It allows for a cost-based systematic control of the trade-off based on the relative importance between the storage and energy in a given network environment and application. Experimental results show that the proposed method achieves a near-optimal control between the storage and energy and reduces the cost by 0.989~84.995 times compared with the cost achieved using the flat (i.e., non-hierarchical) setup as in the work by Xiang et al.

cs.DB↗