Filters








53,952 Hits in 5.6 sec

Learning feedback terms for reactive planning and control

Akshara Rai, Giovanni Sutanto, Stefan Schaal, Franziska Meier
2017 2017 IEEE International Conference on Robotics and Automation (ICRA)  
We approach this by learning a reactive modification term for movement plans represented by nonlinear differential equations.  ...  Reactivity can be accomplished through replanning, e.g. model-predictive control, or through a reactive feedback policy that modifies on-going behavior in response to sensory events.  ...  BACKGROUND We need a representation of planning and control for our work that allows for a flexible insertion of machine learning terms to adapt the planned behavior in response to sensory events.  ... 
doi:10.1109/icra.2017.7989252 dblp:conf/icra/RaiSSM17 fatcat:3nv2mrxdtzb3zkthy5eqau3rsm

Real-time Perception meets Reactive Motion Generation [article]

Daniel Kappler, Jim Mainprice , Vincent Berenz and Jeannette Bohg Autonomous Motion Department at the MPI for Intelligent Systems, Tübingen, Germany, CLMC lab at the University of Southern California, Los Angeles, CA, USA, Lula Robotics Inc., Seattle, WA, USA, Dept. of Computer Science & Engineering, Univ. of Washington (+1 others)
2017 arXiv   pre-print
only reacts to local environment dynamics and (iii) a reactive planner that integrates feedback control and motion optimization.  ...  We also report on the lessons learned for system building.  ...  Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the funding organizations.  ... 
arXiv:1703.03512v3 fatcat:y3y25efvw5fsjeqw6j3xh7glve

Supervised Learning and Reinforcement Learning of Feedback Models for Reactive Behaviors: Tactile Feedback Testbed [article]

Giovanni Sutanto, Katharina Rombach, Yevgen Chebotar, Zhe Su, Stefan Schaal, Gaurav S. Sukhatme, Franziska Meier
2020 arXiv   pre-print
In this paper we introduce a full framework for learning feedback models for reactive motion planning.  ...  In the final phase, a sample-efficient reinforcement learning algorithm fine-tunes these feedback models for novel task settings through few real system interactions.  ...  Moreover, this research was also supported in part by the Max-Planck-Society through funding provided to Giovanni Sutanto, Katharina Rombach, Yevgen Chebotar, Zhe Su, and Stefan Schaal.  ... 
arXiv:2007.00450v1 fatcat:chzm5eduvbe2padut4kedgvfsy

Embodied Artificial Intelligence through Distributed Adaptive Control: An Integrated Framework [article]

Clément Moulin-Frier, Jordi-Ysard Puigbò, Xerxes D. Arsiwalla, Martì Sanchez-Fibla, Paul F. M. J. Verschure
2017 arXiv   pre-print
Regarding integration, we note that the most impactful recent contributions have been made possible through the integration of recent Machine Learning methods (based in particular on Deep Learning and  ...  We then claim that benchmarking environments need to reproduce ecologically-valid conditions for bootstrapping the acquisition of increasingly complex cognitive skills through the concept of a cognitive  ...  agent and its environment and (b) this interaction provides the basis for learning higher-level representations and for sequencing them in a causal way for top-down goal-oriented control.  ... 
arXiv:1704.01407v3 fatcat:jqdvpkzjrnedzpxfz7bxhfn5mm

SOVEREIGN: An autonomous neural system for incrementally learning planned action sequences to navigate towards a rewarded goal

William Gnadt, Stephen Grossberg
2008 Neural Networks  
Controllers for both animals and mobile robots, or animats, need reactive mechanisms for exploration, and learned plans to reach goal objects once an environment becomes familiar.  ...  These working memories trigger learning of sensory and motor sequence categories, or plans, which together control planned movements.  ...  However, the Reactive DV and Planned DV cannot control the animat at the same time.  ... 
doi:10.1016/j.neunet.2007.09.016 pmid:17996419 fatcat:v33zs3egr5edbcc7hjzatzp3ua

SOVEREIGN: an Autonomous Neural System for Incrementally Learning to Navigate towards a Rewarded Goal [chapter]

William Gnadt, Stephen Grossberg
2008 Motion Planning  
The hungry animat learns to categorize the triangle cue and updates the Visual Working Memory and Planning System. It approaches the triangle cue under reactive control.  ...  The Visual and Motor Working Memory and Planning System are updated and the animat approaches the square cue under reactive control.  ...  The book is intended for the readers who are interested and active in the field of robotics and especially for those who want to study and develop their own methods in motion/path planning or control for  ... 
doi:10.5772/6018 fatcat:rv3kg6czzbb7dkjm7q35qhxt7y

Distinct replay signatures for planning and memory maintenance [article]

G. Elliott Wimmer, Yunzhe Liu, Daniel McNamee, Raymond Dolan
2021 bioRxiv   pre-print
Using magnetoencephalography (MEG) and multivariate analysis, we found neural evidence for compressed forward replay during planning and backward replay following reward feedback.  ...  Our reward learning task required human participants to utilize structure knowledge for 'model-based' evaluation, while maintaining knowledge for two independent and randomly alternating task environments  ...  and Krajbich, 2016) and 2) decreased the dependence of reward learning on short-term working memory for preceding feedback (Collins and Frank, 2012; Wimmer et al., 2018) .  ... 
doi:10.1101/2021.11.08.467745 fatcat:bsj4m4iakrgizcql2fej24fdcu

Top-Down and Bottom-Up Interactions between Low-Level Reactive Control and Symbolic Rule Learning in Embodied Agents

Clément Moulin-Frier, Xerxes D. Arsiwalla, Jordi-Ysard Puigbo, Martí Sánchez-Fibla, Armin Duff, Paul F. M. J. Verschure
2016 Neural Information Processing Systems  
We investigate this question in the framework of the "Distributed Adaptive Control" (DAC) cognitive architecture, where we study top-down and bottomup interactions between low-level reactive control and  ...  symbolic rule learning in embodied agents.  ...  • How are those representations recruited in rule and plan learning? • How do rules and plans modulate reactive behaviors through top-down control for realizing longterm goals?  ... 
dblp:conf/nips/Moulin-FrierAPS16 fatcat:zepcxsghffhcvgiigrowskpfee

The Striatum and Subthalamic Nucleus as Independent and Collaborative Structures in Motor Control

Alia Tewari, Rachna Jog, Mandar S. Jog
2016 Frontiers in Systems Neuroscience  
This review compares the role of the striatum and STN in motor response inhibition and execution, competing motor programs, feedback based learning, and response planning.  ...  Research is warranted on the functional connectivity of the network for inhibition involving the rIFG, preSMA, striatum, and STN.  ...  Jog has received honoraria from Abbvie, Merz Pharma, Allergan for speaking engagements and for serving on advisory boards. Dr.  ... 
doi:10.3389/fnsys.2016.00017 pmid:26973474 pmcid:PMC4771745 fatcat:v4buirb2vrbftfntwjqudmrzei

Planning Education for Long-Term Retention: The Cognitive Science and Implementation of Retrieval Practice

Douglas Larsen
2018 Seminars in neurology  
AbstractEducational systems are rarely designed for long-term retention of information.  ...  Repeated acts of retrieval provide opportunities for schemas to be updated and strengthened. Spacing of retrieval allows more consolidated schemas to be reactivated.  ...  Acknowledgment The author would like to thank Andrew Butler for his discussions and insights on the topic, especially his thoughts on consolidation.  ... 
doi:10.1055/s-0038-1666983 pmid:30125899 fatcat:mtb46isqyfbpndlq4kn3ugkva4

Resonant Cholinergic Dynamics in Cognitive and Motor Decision-Making: Attention, Category Learning, and Choice in Neocortex, Superior Colliculus, and Optic Tectum

Stephen Grossberg, Jesse Palma, Massimiliano Versace
2016 Frontiers in Neuroscience  
This map enables visual, auditory, and planned movement commands to compete for attention, leading to selection of a winning position that controls where the next saccadic eye movement will go.  ...  Such task-sensitive control is proposed to control thalamocortical choice of the critical features that are cognitively attended and that are incorporated through learning into prototypes of visual recognition  ...  Gancarz and Grossberg (1999) model how these three processing streams (reactive, attentive, and planned) learn to control accurate saccadic eye movements, despite having different maps and parameters.  ... 
doi:10.3389/fnins.2015.00501 pmid:26834535 pmcid:PMC4718999 fatcat:gajngl6clnbejmjmaqlgtvw55y

Linking perception and action through motivation and affect

Darryl N. Davis
2008 Journal of experimental and theoretical artificial intelligence (Print)  
This research is in effect integrative, drawing together a number of threads in cognitive science and artificial intelligence, for example behaviour, decision-making, memory, and learning.  ...  Cognition involves the control of behaviour both within external environments and internal to an autonomous system.  ...  Importance and Insistence are defined in terms of the affect model. is defined in terms of plans, activity or behavior (for example a reactive subarchitecture specification) that may be suitable for achieving  ... 
doi:10.1080/09528130701472424 fatcat:mujdtvy65zf4bfvosea2ib3usq

Online movement adaptation based on previous sensor experiences

P. Pastor, L. Righetti, M. Kalakrishnan, S. Schaal
2011 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems  
We exploit their dynamic properties by coupling them with the measured and predicted sensor traces. This feedback loop allows for online adaptation of the movement plan.  ...  In this paper, we propose a general framework to achieve very contact-reactive motions for robotic grasping and manipulation.  ...  Second, execute the learned behavior in open loop mode, i.e. setting the feedback terms ζ = 0 in Eq. (9) , and record the experienced internal forces and torques.  ... 
doi:10.1109/iros.2011.6048819 fatcat:z5m2awesc5hdhedqkyqhmcquku

Online movement adaptation based on previous sensor experiences

Peter Pastor, Ludovic Righetti, Mrinal Kalakrishnan, Stefan Schaal
2011 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems  
We exploit their dynamic properties by coupling them with the measured and predicted sensor traces. This feedback loop allows for online adaptation of the movement plan.  ...  In this paper, we propose a general framework to achieve very contact-reactive motions for robotic grasping and manipulation.  ...  Second, execute the learned behavior in open loop mode, i.e. setting the feedback terms ζ = 0 in Eq. (9) , and record the experienced internal forces and torques.  ... 
doi:10.1109/iros.2011.6095059 dblp:conf/iros/PastorRKS11 fatcat:gaagnhssqbgsbbtr2da6fbweda

Gaze behavior when learning to link sequential action phases in a manual task

D. Safstrom, R. S. Johansson, J. R. Flanagan
2014 Journal of Vision  
participants in terms of learning to predictively control the cursor (and hence learn the duration of the hold phase).  ...  to the minimum time for gaze shifts in reactive trials from the reactive control condition ( Figure 3B ).  ...  Stockholm, Sweden, and Canadian Institutes of Health Research.  ... 
doi:10.1167/14.4.3 pmid:24695992 fatcat:czf2hxulbrempjrguzni733yh4
« Previous Showing results 1 — 15 out of 53,952 results