Filters








498,973 Hits in 5.9 sec

Learning the Difference between Partially Observable Dynamical Systems [chapter]

Sami Zhioua, Doina Precup, François Laviolette, Josée Desharnais
2009 Lecture Notes in Computer Science  
We propose a new approach for estimating the difference between two partially observable dynamical systems.  ...  ) between the systems.  ...  The authors thank Prakash Panangaden for helpful discussions.  ... 
doi:10.1007/978-3-642-04174-7_43 fatcat:f2skgwfsefdjdg767kdtnymyky

Learning emergent PDEs in a learned emergent space [article]

Felix P. Kemeth, Tom Bertalan, Thomas Thiem, Felix Dietrich, Sung Joon Moon, Carlo R. Laing, Ioannis G. Kevrekidis
2020 arXiv   pre-print
We extract data-driven, intrinsic spatial coordinates from observations of the dynamics of large systems of coupled heterogeneous agents.  ...  These coordinates then serve as an emergent space in which to learn predictive models in the form of partial differential equations (PDEs) for the collective description of the coupled-agent system.  ...  Acknowledgements This work was partially supported by U.S. Army Research Office (through a MURI program), DARPA, and the U.S. Department of Energy.  ... 
arXiv:2012.12738v1 fatcat:uxxcmejxfnbxjobxioh2o76uka

Learning States Representations in POMDP [article]

Gabriella Contardo and Ludovic Denoyer and Thierry Artieres and Patrick Gallinari
2014 arXiv   pre-print
We propose to deal with sequential processes where only partial observations are available by learning a latent representation space on which policies may be accurately learned.  ...  Acknowledgements This work was performed within the Labex SMART supported by French state funds managed by the ANR within the Investissements d'Avenir programme under reference ANR-11-LABX-65 and by the  ...  This unsupervised operation is used to learn the system only once, and may be used to tackle different tasks sharing the same dynamical process.  ... 
arXiv:1312.6042v4 fatcat:jjbvl5wdmzftjmma577altjqru

Generalizing Over Uncertain Dynamics for Online Trajectory Generation [chapter]

Beomjoon Kim, Albert Kim, Hongkai Dai, Leslie Kaelbling, Tomas Lozano-Perez
2017 Springer Proceedings in Advanced Robotics  
When the dynamics is certain, the algorithm generalizes across model parameters. When the dynamics is partially observable, the algorithm generalizes across different observations.  ...  We present an algorithm which learns an online trajectory generator that can generalize over varying and uncertain dynamics.  ...  We also gratefully acknowledge support from the ONR (grant N00014-14-1-0486 ), from the AFOSR (grant FA23861014135), and from the ARO (grant W911NF1410433).  ... 
doi:10.1007/978-3-319-60916-4_3 dblp:conf/isrr/KimKDKL15 fatcat:kba7ncbr6jbanitdppekvqzrge

Learning second order coupled differential equations that are subject to non-conservative forces [article]

Roger Alexander Müller, Jonathan Laflamme-Janssen, Jaime Camacaro, Carolina Bessega
2021 arXiv   pre-print
even if the system is only partially observed.  ...  In this article we address the question whether it is possible to learn the differential equations describing the physical properties of a dynamical system, subject to non-conservative forces, from observations  ...  In this sense Equation 4 builds the link between the network architecture and the dynamics of the system under observation.  ... 
arXiv:2010.11270v2 fatcat:iwfu7deyenagdjvrzafsl7tddu

Partial Local FriendQ Multiagent Learning: Application to Team Automobile Coordination Problem [chapter]

Julien Laumonier, Brahim Chaib-draa
2006 Lecture Notes in Computer Science  
We compare empirically the performance of the learned policy for totally observable problems and performances of policies for different degrees of observability.  ...  If each degree of observability is associated with communication costs, multiagent system designers are able to choose a compromise between the performance of the policy and the cost to obtain the associated  ...  Also, it will be very interesting to study the effect of partial local view to non cooperative cases.  ... 
doi:10.1007/11766247_31 fatcat:tmentiqzerfdlcj5yhbyg2gdr4

Manifold Embeddings for Model-Based Reinforcement Learning under Partial Observability

Keith Bush, Joelle Pineau
2009 Neural Information Processing Systems  
If partial observability can be overcome, these constraints suggest the use of model-based reinforcement learning.  ...  We demonstrate that the embedding of a system can change as a result of learning, and we argue that the best performing embeddings well-represent the dynamics of both the uncontrolled and adaptively controlled  ...  Acknowledgments The authors thank Dr. Gabriella Panuccio and Dr. Massimo Avoli of the Montreal Neurological Institute for generating the time-series described in Section 4.  ... 
dblp:conf/nips/BushP09 fatcat:orqjthixyvf4lkkwbnknii6v6i

Deep Learning of Chaotic Systems from Partially-Observed Data [article]

Victor Churchill, Dongbin Xiu
2022 arXiv   pre-print
Recently, a general data driven numerical framework has been developed for learning and modeling of unknown dynamical systems using fully- or partially-observed data.  ...  Here we employ several other qualitative and quantitative measures to determine whether the chaotic dynamics have been learned.  ...  systems [35] , parametric dynamical systems [34] , partially observed dynamical systems [13] , as well as partial differential equations [10, 58] .  ... 
arXiv:2205.08384v1 fatcat:p5dp5sahhje75gbe6oca5pmlsq

Unsupervised Graph Neural Network Reveals the Structure–Dynamics Correlation in Disordered Systems [article]

Vaibhav Bihani, Sahil Manchanda, Sayan Ranu, N. M. Anoop Krishnan
2022 arXiv   pre-print
Learning the structure--dynamics correlation in disordered systems is a long-standing problem.  ...  Here, we use unsupervised machine learning employing graph neural networks (GNN) to investigate the local structures in disordered systems.  ...  Except a minor broadening of the first peak of B-B partial PDFs, the partial PDFs of different glassy structures are also comparable.  ... 
arXiv:2206.12575v1 fatcat:wwatz5fnmzdrzpqrgleaddghpq

Prediction and Control in a Dynamic Environment

Magda Osman, Maarten Speekenbrink
2012 Frontiers in Psychology  
The present study compared the accuracy of cue-outcome knowledge gained during prediction-based and control-based learning in stable and unstable dynamic environments.  ...  Participants either learnt to make cue-interventions in order to control an outcome, or learnt to predict the outcome from observing changes to the cue values.  ...  ACKNOWLEDGMENTS The support of the ESRC Research Centre for Economic Learning and Human Evolution is gratefully acknowledged.  ... 
doi:10.3389/fpsyg.2012.00068 pmid:22419913 pmcid:PMC3300109 fatcat:3doeivtlnvg7hlbj7dxl6zv7je

Deep Reinforcement Learning for Dynamic Spectrum Sensing and Aggregation in Multi-Channel Wireless Networks [article]

Yunzeng Li, Wensheng Zhang, Cheng-Xiang Wang, Jian Sun, Yu Liu
2020 arXiv   pre-print
The simulation results show that DQN can achieve near-optimal performance among different system scenarios only based on partial observations and ACK signals.  ...  The performance of DQN, Q-Learning, and the Improvident Policy with known system dynamics is evaluated through simulations.  ...  However, the user is not able to tell the first two situations in practice due to partial observation of the system, but the difference between them is taken into account in simulations to evaluate the  ... 
arXiv:2007.13965v1 fatcat:6ve3stnvffat3j67rzzblifafe

Interaction primitives for human-robot cooperation tasks

Heni Ben Amor, Gerhard Neumann, Sanket Kamthe, Oliver Kroemer, Jan Peters
2014 2014 IEEE International Conference on Robotics and Automation (ICRA)  
Interaction primitives build on the framework of dynamic motor primitives (DMPs) by maintaining a distribution over the parameters of the DMP.  ...  In this paper, we propose to learn interaction skills by observing how two humans engage in a similar task. To this end, we introduce a new representation called Interaction Primitives.  ...  ACKNOWLEDGEMENTS The work presented in this paper is funded through the Daimler-and-Benz Foundation and the European Communitys Seventh Framework Programme under the grant agreement n ICT-600716 (CoDyCo  ... 
doi:10.1109/icra.2014.6907265 dblp:conf/icra/AmorNKKP14 fatcat:jv44caoxtfgbzhjim6xq2rfzka

Variational Deep Learning for the Identification and Reconstruction of Chaotic and Stochastic Dynamical Systems from Noisy and Partial Observations [article]

Duong Nguyen, Said Ouala, Lucas Drumetz, Ronan Fablet
2021 arXiv   pre-print
Within the proposed framework, we jointly learn an inference model to reconstruct the true states of the system and the governing laws of these states from series of noisy and partial data.  ...  However, the identification of governing equations remains challenging when dealing with noisy and partial observations.  ...  Hence, we address here the problem of learning dynamical systems from not only noisy but also partial observations 6 .  ... 
arXiv:2009.02296v6 fatcat:hxpbv65ylbd7vcailahk2cb5u4

Understanding collective crowd behaviors: Learning a Mixture model of Dynamic pedestrian-Agents

Bolei Zhou, Xiaogang Wang, Xiaoou Tang
2012 2012 IEEE Conference on Computer Vision and Pattern Recognition  
Furthermore, MDA can well infer the past behaviors and predict the future behaviors of pedestrians given their trajectories only partially observed, and classify different pedestrian behaviors in the scene  ...  From the agent-based modeling, each pedestrian in the crowd is driven by a dynamic pedestrian-agent, which is a linear dynamic system with its initial and termination states reflecting a pedestrian's belief  ...  We denote D = (A, Γ) as the dynamics parameters to be learned for the agent system.  ... 
doi:10.1109/cvpr.2012.6248013 dblp:conf/cvpr/ZhouWT12 fatcat:divaxytub5fktlgdzvjb54auf4

Physics-Incorporated Convolutional Recurrent Neural Networks for Source Identification and Forecasting of Dynamical Systems [article]

Priyabrata Saha, Saurabh Dash, Saibal Mukhopadhyay
2021 arXiv   pre-print
We formulate our model PhICNet as a convolutional recurrent neural network (RNN) which is end-to-end trainable for spatio-temporal evolution prediction of dynamical systems and learns the source behavior  ...  Spatio-temporal dynamics of physical processes are generally modeled using partial differential equations (PDEs).  ...  Rather, we need to formulate a supervised learning task for the (observable) combined dynamics that will inherently learn the hidden source dynamics.  ... 
arXiv:2004.06243v3 fatcat:ev3u5wf4b5fzbliqpkqojalcga
« Previous Showing results 1 — 15 out of 498,973 results