Filters








4,393 Hits in 5.6 sec

Combining Model-Based and Model-Free Updates for Trajectory-Centric Reinforcement Learning [article]

Yevgen Chebotar, Karol Hausman, Marvin Zhang, Gaurav Sukhatme, Stefan Schaal, Sergey Levine
<span title="2017-06-18">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Reinforcement learning (RL) algorithms for real-world robotic applications need a data-efficient learning process and the ability to handle complex, unknown dynamical systems.  ...  These requirements are handled well by model-based and model-free RL approaches, respectively. In this work, we aim to combine the advantages of these two types of methods in a principled manner.  ...  Acknowledgements The authors would like to thank Sean Mason for his help with preparing the real robot experiments.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.03078v3">arXiv:1703.03078v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/cricu37rongvlkkevecodhszse">fatcat:cricu37rongvlkkevecodhszse</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200914153300/https://arxiv.org/pdf/1703.03078v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a2/e2/a2e2770565665a5d7ba4570934aa1a7882a4e214.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.03078v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Fast deep reinforcement learning using online adjustments from the past [article]

Steven Hansen, Pablo Sprechmann, Alexander Pritzel, André Barreto, Charles Blundell
<span title="2018-10-18">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
EVA combines a number of recent ideas around combining episodic memory-like structures into reinforcement learning agents: slot-based storage, content-based retrieval, and memory-based planning.  ...  We propose Ephemeral Value Adjusments (EVA): a means of allowing deep reinforcement learning agents to rapidly adapt to experience in their replay buffer.  ...  Finally, we thank the anonymous reviewers for their comments and suggestions to improve the paper.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1810.08163v1">arXiv:1810.08163v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ekotbhxuxrafxcc3f4ka4fbvea">fatcat:ekotbhxuxrafxcc3f4ka4fbvea</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200907102429/https://arxiv.org/pdf/1810.08163v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/92/50/9250b29402a3f1b58aef52fc5142e8bcfd077f65.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1810.08163v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Reset-Free Guided Policy Search: Efficient Deep Reinforcement Learning with Stochastic Initial States [article]

William Montgomery, Anurag Ajay, Chelsea Finn, Pieter Abbeel, Sergey Levine
<span title="2016-10-06">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we propose a new reinforcement learning algorithm for learning manipulation skills that can train general-purpose neural network policies with minimal human engineering, while still allowing  ...  for fast, efficient learning in stochastic environments.  ...  One promising direction for future work is to extend this approach to more powerful trajectory-centric RL methods, including stochastic optimization methods [2] and model-free trajectory optimization  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.01112v2">arXiv:1610.01112v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/5t75jq6ctrb3fkg2rg3wfymsoy">fatcat:5t75jq6ctrb3fkg2rg3wfymsoy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200924210826/https://arxiv.org/pdf/1610.01112v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/80/40/8040aa79bda5b3fa8e0845a4298c6a55af155c36.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.01112v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Residual Robot Learning for Object-Centric Probabilistic Movement Primitives [article]

Joao Carvalho, Dorothea Koert, Marek Daniv, Jan Peters
<span title="2022-03-08">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Therefore, we propose to combine ProMPs with recently introduced Residual Reinforcement Learning (RRL), to account for both, corrections in position and orientation during task execution.  ...  It is desirable for future robots to quickly learn new tasks and adapt learned skills to constantly changing environments.  ...  We overcome this problem by proposing to learn a residual policy in position and orientation with model-free reinforcement learning.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.03918v1">arXiv:2203.03918v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/muqcsnntuzdzjk4yefwzu2kw4a">fatcat:muqcsnntuzdzjk4yefwzu2kw4a</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220311161144/https://arxiv.org/pdf/2203.03918v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/f7/72/f772fa5d38c1c29e46e1bd8d77824b82356df9dd.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.03918v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

DEALIO: Data-Efficient Adversarial Learning for Imitation from Observation [article]

Faraz Torabi, Garrett Warnell, Peter Stone
<span title="2021-03-31">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
, model-free reinforcement learning algorithms.  ...  In this work, we hypothesize that we can incorporate ideas from model-based reinforcement learning with adversarial methods for IfO in order to increase the data efficiency of these methods without sacrificing  ...  Reinforcement Learning Broadly speaking, reinforcement learning algorithms can be categorized as model-based or model-free.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2104.00163v1">arXiv:2104.00163v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/z4vjlyf7snbhznd2ro5z7wfyyu">fatcat:z4vjlyf7snbhznd2ro5z7wfyyu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210406033736/https://arxiv.org/pdf/2104.00163v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/33/0b/330bd0ae3926e294f728f29dc5510ad51f81d427.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2104.00163v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Reinforcement and Imitation Learning for Diverse Visuomotor Skills [article]

Yuke Zhu, Ziyu Wang, Josh Merel, Andrei Rusu, Tom Erez, Serkan Cabi, Saran Tunyasuvunakool, János Kramár, Raia Hadsell, Nando de Freitas, Nicolas Heess
<span title="2018-05-27">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We propose a model-free deep reinforcement learning method that leverages a small amount of demonstration data to assist a reinforcement learning agent.  ...  In experiments, our reinforcement and imitation agent achieves significantly better performances than agents trained with reinforcement learning or imitation learning alone.  ...  ACKNOWLEDGMENT The authors would like to thank Yuval Tassa, Jonathan Scholz, Thomas Rothörl, Jonathan Hunt, and many other colleagues at DeepMind for the helpful discussion and feedback.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1802.09564v2">arXiv:1802.09564v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/7vwziswy25blbdgmb6gklibl5m">fatcat:7vwziswy25blbdgmb6gklibl5m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191020130739/https://arxiv.org/pdf/1802.09564v1.pdf" title="fulltext PDF download [not primary version]" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <span style="color: #f43e3e;">&#10033;</span> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/56/f7/56f73c1d42501dd5f5d559f8649781e6f95b9bba.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1802.09564v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Reinforcement and Imitation Learning for Diverse Visuomotor Skills

Yuke Zhu, Ziyu Wang, Josh Merel, Andrei Rusu, Tom Erez, Serkan Cabi, Saran Tunyasuvunakool, János Kramár, Raia Hadsell, Nando de Freitas, Nicolas Heess
<span title="2018-06-26">2018</span> <i title="Robotics: Science and Systems Foundation"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/gjhqqq6dgnaupkvp2ckhefvv6i" style="color: black;">Robotics: Science and Systems XIV</a> </i> &nbsp;
We propose a model-free deep reinforcement learning method that leverages a small amount of demonstration data to assist a reinforcement learning agent.  ...  In experiments, our reinforcement and imitation agent achieves significantly better performances than agents trained with reinforcement learning or imitation learning alone.  ...  ACKNOWLEDGMENT The authors would like to thank Yuval Tassa, Jonathan Scholz, Thomas Rothörl, Jonathan Hunt, and many other colleagues at DeepMind for the helpful discussion and feedback.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.15607/rss.2018.xiv.009">doi:10.15607/rss.2018.xiv.009</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/rss/Zhu0MRECTKHFH18.html">dblp:conf/rss/Zhu0MRECTKHFH18</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/u6pt5wi6lvgchhezcsn6qntid4">fatcat:u6pt5wi6lvgchhezcsn6qntid4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190225131533/http://pdfs.semanticscholar.org/6318/ffe9649e4c4d683da2336f57c1a45189a692.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/63/18/6318ffe9649e4c4d683da2336f57c1a45189a692.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.15607/rss.2018.xiv.009"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / doi.org </button> </a>

Learning Dexterous Grasping with Object-Centric Visual Affordances [article]

Priyanka Mandikal, Kristen Grauman
<span title="2021-06-16">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Unlike traditional approaches that learn from human demonstration trajectories (e.g., hand joint sequences captured with a glove), the proposed prior is object-centric and image-based, allowing the agent  ...  Our key idea is to embed an object-centric visual affordance model within a deep reinforcement learning loop to learn grasping policies that favor the same object regions favored by people.  ...  Towards this end, we develop a deep model-free reinforcement learning model for dexterous grasping. Our robot model assumes access to visual sensing and proprioception, as well as 3D point tracking.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2009.01439v2">arXiv:2009.01439v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/xvlhvzxokjcshnlv6zmmo6kece">fatcat:xvlhvzxokjcshnlv6zmmo6kece</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210620023340/https://arxiv.org/pdf/2009.01439v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/5f/6e/5f6e334abecb5f6e0ec34c2c38b13f4aa8913bce.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2009.01439v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations [article]

Aravind Rajeswaran, Vikash Kumar, Abhishek Gupta, Giulia Vezzani, John Schulman, Emanuel Todorov, Sergey Levine
<span title="2018-06-26">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Deep reinforcement learning (DRL) provides a model-agnostic approach to control complex dynamical systems, but has not been shown to scale to high-dimensional dexterous manipulation.  ...  In this work, we show that model-free DRL can effectively scale up to complex manipulation tasks with a high-dimensional 24-DoF hand, and solve them from scratch in simulated experiments.  ...  ACKNOWLEDGEMENTS The authors would like to thank Ilya Sutskever, Wojciech Zaremba, Igor Mordatch, Pieter Abbeel, Ankur Handa, Oleg Klimov, Sham Kakade, Ashvin Nair, and Kendall Lowrey for valuable comments  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1709.10087v2">arXiv:1709.10087v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/jauhf3xiajginlter6jyjr2iq4">fatcat:jauhf3xiajginlter6jyjr2iq4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200829031650/https://arxiv.org/pdf/1709.10087v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/aa/74/aa746af3a6f19bcc004265bbe3793e001e4793c4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1709.10087v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations

Aravind Rajeswaran, Vikash Kumar, Abhishek Gupta, Giulia Vezzani, John Schulman, Emanuel Todorov, Sergey Levine
<span title="2018-06-26">2018</span> <i title="Robotics: Science and Systems Foundation"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/gjhqqq6dgnaupkvp2ckhefvv6i" style="color: black;">Robotics: Science and Systems XIV</a> </i> &nbsp;
Deep reinforcement learning (DRL) provides a model-agnostic approach to control complex dynamical systems, but has not been shown to scale to highdimensional dexterous manipulation.  ...  In this work, we show that model-free DRL can effectively scale up to complex manipulation tasks with a high-dimensional 24-DoF hand, and solve them from scratch in simulated experiments.  ...  ACKNOWLEDGEMENTS The authors would like to thank Ilya Sutskever, Wojciech Zaremba, Igor Mordatch, Pieter Abbeel, Ankur Handa, Oleg Klimov, Sham Kakade, Ashvin Nair, and Kendall Lowrey for valuable comments  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.15607/rss.2018.xiv.049">doi:10.15607/rss.2018.xiv.049</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/rss/RajeswaranKGVST18.html">dblp:conf/rss/RajeswaranKGVST18</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ncr5zhal7ja4hc4yiz5ruhxdau">fatcat:ncr5zhal7ja4hc4yiz5ruhxdau</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190228203702/http://pdfs.semanticscholar.org/980c/2f4452ac18272f41ce1a9e2e9a8419f740f4.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/98/0c/980c2f4452ac18272f41ce1a9e2e9a8419f740f4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.15607/rss.2018.xiv.049"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / doi.org </button> </a>

Collective Robot Reinforcement Learning with Distributed Asynchronous Guided Policy Search [article]

Ali Yahya, Adrian Li, Mrinal Kalakrishnan, Yevgen Chebotar, Sergey Levine
<span title="2016-10-03">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In principle, reinforcement learning and policy search methods can enable robots to learn highly complex and general skills that may allow them to function amid the complexity and diversity of the real  ...  Fortunately, it is possible for multiple robots to share their experience with one another, and thereby, learn a policy collectively.  ...  ACKNOWLEDGEMENTS We would like to thank Peter Pastor and Kurt Konolige for additional engineering, robot maintenance, and technical discussions, and Ryan Walker and Gary Vosters for designing custom hardware  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.00673v1">arXiv:1610.00673v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/iyhfhs6mwbaktp3yg3wcixr6yq">fatcat:iyhfhs6mwbaktp3yg3wcixr6yq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200929183150/https://arxiv.org/pdf/1610.00673v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/03/34/0334cf4857b465184f969c65f1ea0347d452ff18.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.00673v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Unsupervised Visuomotor Control through Distributional Planning Networks [article]

Tianhe Yu, Gleb Shevchuk, Dorsa Sadigh, Chelsea Finn
<span title="2019-02-14">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
This enables learning effective and control-centric representations that lead to more autonomous reinforcement learning algorithms.  ...  for autonomous reinforcement learning.  ...  ACKNOWLEDGMENTS We thank Aravind Srinivas and Sergey Levine for helpful discussions.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.05542v1">arXiv:1902.05542v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/eqcmujyykjfyhgxkvu7z7y3wxy">fatcat:eqcmujyykjfyhgxkvu7z7y3wxy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191027055621/https://arxiv.org/pdf/1902.05542v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/ef/ce/efce87c9d5cc314d0c3c53286b9d2bbed7fd8775.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.05542v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

GrASP: Gradient-Based Affordance Selection for Planning [article]

Vivek Veeriah, Zeyu Zheng, Richard Lewis, Satinder Singh
<span title="2022-02-08">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
model can outperform model-free RL.  ...  Planning with a learned model is arguably a key component of intelligence. There are several challenges in realizing such a component in large-scale reinforcement learning (RL) problems.  ...  Gradient through a trajectory in a model to update policies.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2202.04772v1">arXiv:2202.04772v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/gsdeyvfnqjdrbgt3tcqssfdo34">fatcat:gsdeyvfnqjdrbgt3tcqssfdo34</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220212060242/https://arxiv.org/pdf/2202.04772v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/e6/17/e617b03e14bbdde9f0fb3f9fabca0dd29c91f174.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2202.04772v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Path Integral Guided Policy Search [article]

Yevgen Chebotar, Mrinal Kalakrishnan, Ali Yahya, Adrian Li, Stefan Schaal, Sergey Levine
<span title="2018-10-11">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We extend GPS in the following ways: (1) we propose the use of a model-free local optimizer based on path integral stochastic optimal control (PI2), which enables us to learn local policies for tasks with  ...  We validate the method on a challenging door opening task and a pick-and-place task, and we demonstrate that our approach substantially outperforms the prior LQR-based local policy optimizer on these tasks  ...  ACKNOWLEDGEMENTS We would like to thank Peter Pastor and Kurt Konolige for additional engineering, robot maintenance, and technical discussions, and Ryan Walker and Gary Vosters for designing custom hardware  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.00529v2">arXiv:1610.00529v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/tji5loidmbhhbbv3szc4vftuci">fatcat:tji5loidmbhhbbv3szc4vftuci</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191014180230/https://arxiv.org/pdf/1610.00529v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/47/1f/471f0b84932853a6b03a504b65be45554a9a9520.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1610.00529v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Model Mediated Teleoperation with a Hand-Arm Exoskeleton in Long Time Delays Using Reinforcement Learning [article]

Hadi Beik-Mohammadi, Matthias Kerzel, Benedikt Pleintinger, Thomas Hulin, Philipp Reisich, Annika Schmidt, Aaron Pereira, Stefan Wermter, Neal Y. Lii
<span title="2021-07-01">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Augmented reality was also provided to fuse the avatar device and virtual environment models for the teleoperator.  ...  As one of the best alternatives to human-level intelligence, Reinforcement Learning (RL) may offer a solution to cope with these issues.  ...  Our approach adopts the architecture from the MMT scheme and combines it with Dynamic Movement Primitives (DMP) [28] and model-free Reinforcement Learning (RL) [6] , [29] , [30] .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.00359v1">arXiv:2107.00359v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/dvrjless7zeb7ezrjetdy66igm">fatcat:dvrjless7zeb7ezrjetdy66igm</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210707004746/https://arxiv.org/pdf/2107.00359v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/65/b9/65b93881439a8a5fdfecc63bf911add6dcd9853a.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.00359v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 4,393 results