no code implementations • 20 May 2024 • Yiqing Xu, Jiayuan Mao, Yilun Du, Tomas Lozáno-Pérez, Leslie Pack Kaebling, David Hsu
This paper studies the challenge of developing robots capable of understanding under-specified instructions for creating functional object arrangements, such as "set up a dining table for two"; previous arrangement approaches have focused on much more explicit instructions, such as "put object A on the table."
no code implementations • 17 Apr 2024 • Liwei Kang, Zirui Zhao, David Hsu, Wee Sun Lee
We found that if problems can be decomposed into a sequence of reasoning steps and learning to predict the next step has a low sample and computational complexity, explicitly outlining the reasoning chain with all necessary information for predicting the next step may improve performance.
no code implementations • 7 Jan 2024 • Connie Jiang, Yiqing Xu, David Hsu
The advantages of pre-trained large language models (LLMs) are apparent in a variety of language processing tasks.
no code implementations • 29 Nov 2023 • Siwei Chen, Anxing Xiao, David Hsu
We propose an open state representation that provides continuous expansion and updating of object attributes from the LLM's inherent capabilities for context understanding and historical action reasoning.
no code implementations • 23 Oct 2023 • Bo Ai, Zhanxin Wu, David Hsu
The data-driven approach to robot control has been gathering pace rapidly, yet generalization to unseen task domains remains a critical challenge.
no code implementations • 21 Jul 2023 • Yiqing Xu, David Hsu
Tidying up a messy table may appear simple for humans, but articulating clear criteria for tidiness is challenging due to the ambiguous nature of common sense reasoning.
no code implementations • 13 Jul 2023 • Yiqing Xu, Finale Doshi-Velez, David Hsu
Inverse reinforcement learning (IRL) algorithms often rely on (forward) reinforcement learning or planning over a given time horizon to compute an approximately optimal policy for a hypothesized reward function and then match this policy with expert demonstrations.
no code implementations • 12 Mar 2023 • Yiyuan Lee, Katie Lee, Panpan Cai, David Hsu, Lydia E. Kavraki
Identifying internal parameters for planning is crucial to maximizing the performance of a planner.
no code implementations • 1 Oct 2022 • Zirui Zhao, Wee Sun Lee, David Hsu
Natural language generally describes objects and spatial relations with compositionality and ambiguity, two major obstacles to effective language grounding.
1 code implementation • 23 Sep 2022 • Mohamad H. Danesh, Panpan Cai, David Hsu
To address this, we propose a new algorithm, LEarning Attention over Driving bEhavioRs (LEADER), that learns to attend to critical human behaviors during planning.
no code implementations • 21 Sep 2022 • Mikko Lauri, David Hsu, Joni Pajarinen
Noisy sensing, imperfect control, and environment changes are defining characteristics of many real-world robot tasks.
no code implementations • 9 Jun 2022 • Yiqing Xu, Wei Gao, David Hsu
Inverse reinforcement learning (IRL) seeks to infer a cost function that explains the underlying goals and preferences of expert demonstrations.
no code implementations • 25 Feb 2022 • Wei Gao, David Hsu, Wee Sun Lee
To solve these issues, we present Context Hierarchy IRL(CHIRL), a new IRL algorithm that exploits the context to scale up IRL and learn reward functions of complex behaviors.
no code implementations • 16 Sep 2021 • Bo Ai, Wei Gao, Vinay, David Hsu
Importantly, we integrate the multiple neural network modules into a unified controller that achieves robust performance for visual navigation in complex, partially observable environments.
no code implementations • 25 Aug 2021 • Hanbo Zhang, Yunfan Lu, Cunjun Yu, David Hsu, Xuguang Lan, Nanning Zheng
This paper presents INVIGORATE, a robot system that interacts with human through natural language and grasps a specified object in clutter.
no code implementations • 19 Jul 2021 • Siwei Chen, Xiao Ma, Yunfan Lu, David Hsu
Like the model-based analytic approaches to manipulation, the particle representation enables the robot to reason about the object's geometry and dynamics in order to choose suitable manipulation actions.
no code implementations • CVPR 2021 • Peter Karkus, Shaojun Cai, David Hsu
We introduce the Differentiable SLAM Network (SLAM-net) along with a navigation architecture to enable planar robot navigation in previously unseen indoor environments.
no code implementations • 25 Apr 2021 • Xiao Ma, David Hsu, Wee Sun Lee
Manipulating deformable objects, such as ropes and clothing, is a long-standing challenge in robotics, because of their large degrees of freedom, complex non-linear dynamics, and self-occlusion in visual perception.
no code implementations • 11 Jan 2021 • Panpan Cai, David Hsu
To achieve real-time performance for large-scale planning, this work introduces a new algorithm Learning from Tree Search for Driving (LeTS-Drive), which integrates planning and learning in a closed loop, and applies it to autonomous driving in crowded urban traffic in simulation.
Autonomous Driving Robotics
1 code implementation • 7 Nov 2020 • Yiyuan Lee, Panpan Cai, David Hsu
The partially observable Markov decision process (POMDP) is a principled general framework for robot decision making under uncertainty, but POMDP planning suffers from high computational complexity, when long-term planning is required.
1 code implementation • 6 Aug 2020 • Xiao Ma, Siwei Chen, David Hsu, Wee Sun Lee
This paper presents Contrastive Variational Reinforcement Learning (CVRL), a model-based method that tackles complex visual observations in DRL.
1 code implementation • 13 Jul 2020 • Siwei Chen, Xiao Ma, David Hsu
It has been arduous to assess the progress of a policy learning algorithm in the domain of hierarchical task with high dimensional action space due to the lack of a commonly accepted benchmark.
1 code implementation • ICLR 2020 • Xiao Ma, Peter Karkus, David Hsu, Wee Sun Lee, Nan Ye
The particle filter maintains a belief using learned discriminative update, which is trained end-to-end for decision making.
3 code implementations • 11 Nov 2019 • Panpan Cai, Yiyuan Lee, Yuanfu Luo, David Hsu
Autonomous driving in an unregulated urban crowd is an outstanding challenge, especially, in the presence of many aggressive, high-speed traffic participants.
Robotics Multiagent Systems
no code implementations • 3 Sep 2019 • Yaqi Xie, Indu P Bodala, Desmond C. Ong, David Hsu, Harold Soh
In this paper, we present results from a human-subject study designed to explore two facets of human mental models of robots---inferred capability and intention---and their relationship to overall trust and eventual decisions.
1 code implementation • 29 Jul 2019 • Hanbo Zhang, Site Bai, Xuguang Lan, David Hsu, Nanning Zheng
We propose \emph{Hindsight Trust Region Policy Optimization}(HTRPO), a new RL algorithm that extends the highly successful TRPO algorithm with \emph{hindsight} to tackle the challenge of sparse rewards.
1 code implementation • 4 Jun 2019 • Yuanfu Luo, Panpan Cai, Yiyuan Lee, David Hsu
Further, the computational efficiency and the flexibility of GAMMA enable (i) simulation of mixed urban traffic at many locations worldwide and (ii) planning for autonomous driving in dense traffic with uncertain driver behaviors, both in real-time.
1 code implementation • 30 May 2019 • Xiao Ma, Peter Karkus, David Hsu, Wee Sun Lee
Recurrent neural networks (RNNs) have been extraordinarily successful for prediction with sequential data.
no code implementations • 29 May 2019 • Panpan Cai, Yuanfu Luo, Aseem Saxena, David Hsu, Wee Sun Lee
LeTS-Drive leverages the robustness of planning and the runtime efficiency of learning to enhance the performance of both.
no code implementations • 28 May 2019 • Peter Karkus, Xiao Ma, David Hsu, Leslie Pack Kaelbling, Wee Sun Lee, Tomas Lozano-Perez
This paper introduces the Differentiable Algorithm Network (DAN), a composable architecture for robot learning systems.
no code implementations • 26 Apr 2019 • Robert Pinsler, Peter Karkus, Andras Kupcsik, David Hsu, Wee Sun Lee
Our key observation is that experience can be directly generalized over target contexts.
no code implementations • 17 Jul 2018 • Peter Karkus, David Hsu, Wee Sun Lee
We propose to take a novel approach to robot system design where each building block of a larger system is represented as a differentiable program, i. e. a deep neural network.
1 code implementation • Robotics: Science and Systems 2018 • Jue Kun Li, David Hsu, Wee Sun Lee
This paper introduces Push-Net, a deep recurrent neural network model, which enables a robot to push objects of unknown physical properties for re-positioning and re-orientation, using only visual camera images as input.
no code implementations • 11 Jun 2018 • Mohit Shridhar, David Hsu
The first stage uses a neural network to generate visual descriptions of objects, compares them with the input language expression, and identifies a set of candidate objects.
no code implementations • CVPR 2018 • Ziquan Lan, David Hsu, Gim Hee Lee
The user, instead of holding a camera in hand and manually searching for a viewpoint, will interact directly with image contents in the viewfinder through simple gestures, and the flying camera will achieve the desired viewpoint through the autonomous flying capability of the drone.
no code implementations • 30 May 2018 • Yuanfu Luo, Panpan Cai, Aniket Bera, David Hsu, Wee Sun Lee, Dinesh Manocha
Our planning system combines a POMDP algorithm with the pedestrian motion model and runs in near real time.
Robotics
2 code implementations • 23 May 2018 • Peter Karkus, David Hsu, Wee Sun Lee
Particle filtering is a powerful approach to sequential state estimation and finds application in many domains, including robot localization, object tracking, etc.
1 code implementation • 17 Feb 2018 • Panpan Cai, Yuanfu Luo, David Hsu, Wee Sun Lee
Planning under uncertainty is critical for robust robot performance in uncertain, dynamic environments, but it incurs high computational cost.
no code implementations • 12 Jan 2018 • Min Chen, Stefanos Nikolaidis, Harold Soh, David Hsu, Siddhartha Srinivasa
The trust-POMDP model provides a principled approach for the robot to (i) infer the trust of a human teammate through interaction, (ii) reason about the effect of its own actions on human trust, and (iii) choose actions that maximize team performance over the long term.
2 code implementations • 16 Oct 2017 • Wei Gao, David Hsu, Wee Sun Lee, ShengMei Shen, Karthikk Subramanian
How can a delivery robot navigate reliably to a destination in a new office building, with minimal prior information?
no code implementations • 18 Jul 2017 • Mohit Shridhar, David Hsu
A core issue for the system is semantic and spatial grounding, which is to infer objects and their spatial relationships from images and natural language expressions.
2 code implementations • NeurIPS 2017 • Peter Karkus, David Hsu, Wee Sun Lee
It is a recurrent policy network, but it represents a policy for a parameterized set of tasks by connecting a model with a planning algorithm that solves the model, thus embedding the solution structure of planning in a network learning architecture.
no code implementations • 6 Dec 2016 • Peter Karkus, Andras Kupcsik, David Hsu, Wee Sun Lee
Scarce data is a major challenge to scaling robot learning to truly complex tasks, as we need to generalize locally learned policies over different "contexts".
1 code implementation • NeurIPS 2013 • Nan Ye, Adhiraj Somani, David Hsu, Wee Sun Lee
We show that the best policy obtained from a DESPOT is near-optimal, with a regret bound that depends on the representation size of the optimal policy.
no code implementations • 16 Feb 2016 • Min Chen, Emilio Frazzoli, David Hsu, Wee Sun Lee
We show that a POMDP-lite is equivalent to a set of fully observable Markov decision processes indexed by a hidden parameter and is useful for modeling a variety of interesting robotic tasks.
no code implementations • NeurIPS 2015 • Zhan Wei Lim, David Hsu, Wee Sun Lee
Adaptive stochastic optimization optimizes an objective function adaptively under uncertainty.
no code implementations • 6 Nov 2013 • Xinxi Wang, Yi Wang, David Hsu, Ye Wang
Current music recommender systems typically act in a greedy fashion by recommending songs with the highest user ratings.
no code implementations • 27 Jun 2012 • Yi Wang, Kok Sung Won, David Hsu, Wee Sun Lee
Bayesian reinforcement learning (BRL) encodes prior knowledge of the world in a model and represents uncertainty in model parameters by maintaining a probability distribution over them.
no code implementations • NeurIPS 2011 • Zhan Lim, Lee Sun, David Hsu
The recently introduced Monte Carlo Value Iteration (MCVI) can tackle POMDPs with very large discrete state spaces or continuous state spaces, but its performance degrades when faced with long planning horizons.