Filters








87,985 Hits in 7.2 sec

Adapting Reinforcement Learning for Trust: Effective Modeling in Dynamic Environments

Özgür Kafali, Pinar Yolum
2009 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology  
In open multiagent systems, agents need to model their environments in order to identify trustworthy agents.  ...  Contrary to existing modeling approaches that require domain knowledge to build models, our proposed approach can be effectively realized in multiagent systems when the agent's actions are clearly identified  ...  To decide which action is better for it, it uses reinforcement learning. Typically, agents utilizing reinforcement learning model their environments through trial and error interactions [4] .  ... 
doi:10.1109/wi-iat.2009.67 dblp:conf/webi/KafaliY09 fatcat:7ijoprvwufbtlcx7bzo3tvxxy4

Fast Adaptation with Meta-Reinforcement Learning for Trust Modelling in Human-Robot Interaction

Yuan Gao, Elena Sibirtseva, Ginevra Castellano, Danica Kragic
2019 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)  
We present a metalearning based policy gradient method for addressing the problem of adaptation in human-robot interaction and also investigate its role as a mechanism for trust modelling.  ...  We found that our proposed model increased the perceived trustworthiness of the robot and influenced the dynamics of gaining human's trust.  ...  Fast Adaptation with Meta-Reinforcement Learning for Trust Modelling in Human-Robot Interaction Yuan Gao 1 * , Elena Sibirtseva 2 * , Ginevra Castellano 1 and Danica Kragic 2 Abstract-In socially assistive  ... 
doi:10.1109/iros40897.2019.8967924 dblp:conf/iros/GaoSCK19 fatcat:6hfshakl6na37chn6z3wnpe6wq

Sub-policy Adaptation for Hierarchical Reinforcement Learning [article]

Alexander C. Li, Carlos Florensa, Ignasi Clavera, Pieter Abbeel
2020 arXiv   pre-print
Hierarchical reinforcement learning is a promising approach to tackle long-horizon decision-making problems with sparse rewards.  ...  Second, we propose a method for training time-abstractions that improves the robustness of the obtained skills to environment changes. Code and results are available at sites.google.com/view/hippo-rl  ...  ROBUSTNESS TO DYNAMICS PERTURBATIONS We investigate the robustness of HiPPO to changes in the dynamics of the environment.  ... 
arXiv:1906.05862v4 fatcat:vyxyt36vfjeuppbhwaiiwsg4wy

Fast Design Space Adaptation with Deep Reinforcement Learning for Analog Circuit Sizing [article]

Kai-En Yang, Chia-Yu Tsai, Hung-Hao Shen, Chen-Feng Chiang, Feng-Ming Tsai, Chung-An Wang, Yiju Ting, Chia-Shun Yeh, Chin-Tang Lai
2020 arXiv   pre-print
We present a novel framework for design space search on analog circuit sizing using deep reinforcement learning (DRL).  ...  Thus, a simple feed-forward network with few layers can be used to implement a model-based reinforcement learning agent.  ...  An akin method is meta-learning [24] . The model attempts to adapt to new tasks quickly, rather than focusing on a specific environment.  ... 
arXiv:2009.13772v3 fatcat:iy3wk7txpbhctotreakhpb43g4

Terminal Adaptive Guidance for Autonomous Hypersonic Strike Weapons via Reinforcement Learning [article]

Brian Gaudet, Roberto Furfaro
2021 arXiv   pre-print
An adaptive guidance system suitable for the terminal phase trajectory of a hypersonic strike weapon is optimized using reinforcement meta learning.  ...  Finally, we include preliminary results for an integrated guidance and control system in a six degrees-of-freedom environment.  ...  Background: Reinforcement Learning Framework In the reinforcement learning framework, an agent learns through episodic interaction with an environment how to successfully complete a task using a policy  ... 
arXiv:2110.00634v1 fatcat:cbbdebnx7zg6xhj7gx4bj2duym

Deep Reinforcement Learning Empowered Adaptivity for Future Blockchain Networks

Chao Qiu, Xiaoxu Ren, Yifan Cao, Tianle Mai
2020 IEEE Open Journal of the Computer Society  
In this paper, we study a deep reinforcement learning empowered adaptivity approach for future blockchain networks, so as to improve the scalability and meet the requirements of different users.  ...  The deep reinforcement learning empowered adaptivity can help the blockchain network break through the bottleneck.  ...  For a comprehensive perspective, we present the dueling deep reinforcement learning approach in Fig. 3 .  ... 
doi:10.1109/ojcs.2020.3010987 fatcat:w6qpvflmhben5k5qofm3omg5pa

Comparison of Deep Reinforcement Learning and Model Predictive Control for Adaptive Cruise Control [article]

Yuan Lin, John McPhee, Nasser L. Azad
2020 arXiv   pre-print
This study compares Deep Reinforcement Learning (DRL) and Model Predictive Control (MPC) for Adaptive Cruise Control (ACC) design in car-following scenarios.  ...  A first-order system is used as the Control-Oriented Model (COM) to approximate the acceleration command dynamics of a vehicle.  ...  ACKNOWLEDGMENT The authors would like to thank Toyota, Ontario Centres of Excellence, and the Natural Sciences and Engineering Research Council of Canada for the support of this work.  ... 
arXiv:1910.12047v3 fatcat:qosqvikdjrgg7asfk5deivj6iu

CARL: A Benchmark for Contextual and Adaptive Reinforcement Learning [article]

Carolin Benjamins, Theresa Eimer, Frederik Schubert, André Biedenkapp, Bodo Rosenhahn, Frank Hutter, Marius Lindauer
2021 arXiv   pre-print
While Reinforcement Learning has made great strides towards solving ever more complicated tasks, many algorithms are still brittle to even slight changes in their environment.  ...  We show the urgent need of such benchmarks by demonstrating that even simple toy environments become challenging for commonly used approaches if different contextual instances of this task have to be considered  ...  D.3 Challenge III: Continual Learning With the flexibility and easy modifiability of CARLs provided contexts, CARL is suitable for studying continual reinforcement learning agents.  ... 
arXiv:2110.02102v2 fatcat:7hzgrg5puzg5jeyq5uvfyq4cie

Reinforcement Learning Techniques for Decentralized Self-adaptive Service Assembly [chapter]

M. Caporuscio, M. D'Angelo, V. Grassi, R. Mirandola
2016 Lecture Notes in Computer Science  
[4] presented a reinforcement learning model for adaptive resource allocation in a multi-agent system. The learning scheme is based on minority games on networks.  ...  [10] apply reinforcement learning for the dynamic load balancing of parallel data-intensive applications.  ... 
doi:10.1007/978-3-319-44482-6_4 fatcat:i3v26qa4ybfgvkfgrvecwfukmu

Meta Reinforcement Learning for Adaptive Control: An Offline Approach [article]

Daniel G. McClement, Nathan P. Lawrence, Johan U. Backstrom, Philip D. Loewen, Michael G. Forbes, R. Bhushan Gopaluni
2022 arXiv   pre-print
This end-to-end architecture enables the agent to automatically adapt to changes in the process dynamics.  ...  A key design element is the ability to leverage model-based information offline during training, while maintaining a model-free policy structure for interacting with novel environments.  ...  Introduction Reinforcement learning (RL) is a branch of machine learning that formulates a goal-oriented "policy" for taking actions in a stochastic environment [1] .  ... 
arXiv:2203.09661v1 fatcat:ag666udwsvbavixrkmtdxt6gpm

Meta-Adversarial Inverse Reinforcement Learning for Decision-making Tasks [article]

Pin Wang, Hanhan Li, Ching-Yao Chan
2021 arXiv   pre-print
In this work, we build an adaptable imitation learning model based on the integration of Meta-learning and Adversarial Inverse Reinforcement Learning (Meta-AIRL).  ...  Therefore, it is desirable for the trained model to adapt to new tasks that have limited data samples available.  ...  It can be intractable to model the distribution of tasks for highly dynamic environments such as a driving domain.  ... 
arXiv:2103.12694v2 fatcat:mer2u7uplbeuhml2yszq5vco4m

Deep Reinforcement Learning for Single-Shot Diagnosis and Adaptation in Damaged Robots [article]

Shresth Verma, Haritha S. Nair, Gaurav Agarwal, Joydip Dhar, Anupam Shukla
2019 arXiv   pre-print
We thus propose a damage aware control architecture which diagnoses the damage prior to gait selection while also incorporating domain randomization in the damage space for learning a robust policy.  ...  To enable robotic adaptation in such situations, the agent needs to adopt policies which are robust to a diverse set of damages and must do so with minimum computational complexity.  ...  Deep Reinforcement learning (Deep RL) has been shown to be effective in modeling such navigation problems because of both its online and offline learning capabilities in high dimensional search spaces  ... 
arXiv:1910.01240v1 fatcat:adphlgv7hnfldjw6pmxltb4mj4

Distributed Deep Reinforcement Learning for Adaptive Medium Access and Modulation in Shared Spectrum [article]

Akash Doshi, Jeffrey G. Andrews
2022 arXiv   pre-print
We formulate and develop novel distributed implementations of two deep reinforcement learning approaches - Deep Q Networks and Proximal Policy Optimization - modelled on a two stage Markov decision process  ...  Spectrum scarcity has led to growth in the use of unlicensed spectrum for cellular systems.  ...  Related Work Several papers have attempted to utilize reinforcement learning to design a rule for choosing a transmit MCS.  ... 
arXiv:2109.11723v2 fatcat:yb2n3g24ybf77ono2f6zla5viq

Adaptive Quadruped Balance Control for Dynamic Environments Using Maximum-Entropy Reinforcement Learning

Haoran Sun, Tingting Fu, Yuanhuai Ling, Chaoming He
2021 Sensors  
External disturbance poses the primary threat to robot balance in dynamic environments.  ...  This paper provides a learning-based control architecture for quadrupedal self-balancing, which is adaptable to multiple unpredictable scenes of external continuous disturbance.  ...  In our method, the learned balance control policy can be obtained by training in a simulator using an animated dynamic environment and simplified virtual robot models.  ... 
doi:10.3390/s21175907 pmid:34502796 fatcat:sih2ddbxyzayjde3sc3z4zgtui

Learning to Run challenge solutions: Adapting reinforcement learning methods for neuromusculoskeletal environments [article]

Łukasz Kidziński, Sharada Prasanna Mohanty, Carmichael Ong, Zhewei Huang, Shuchang Zhou, Anton Pechenko, Adam Stelmaszczyk, Piotr Jarosik, Mikhail Pavlov, Sergey Kolesnikov, Sergey Plis, Zhibo Chen, Zhizheng Zhang (+16 others)
2018 arXiv   pre-print
In the NIPS 2017 Learning to Run challenge, participants were tasked with building a controller for a musculoskeletal model to make it run as fast as possible through an obstacle course.  ...  In this work, we present eight solutions that used deep reinforcement learning approaches, based on algorithms such as Deep Deterministic Policy Gradient, Proximal Policy Optimization, and Trust Region  ...  In a way, this work is related to the ideas of hierarchical reinforcement learning [10] and the work on learning Dynamic Movement Primitives [30, 13] which are attractor systems of a lower dimensionality  ... 
arXiv:1804.00361v1 fatcat:6r6tklvcmzai5kj37p7gy27h2i
« Previous Showing results 1 — 15 out of 87,985 results