Filters








250,070 Hits in 5.1 sec

Combining Reward Information from Multiple Sources [article]

Dmitrii Krasheninnikov, Rohin Shah, Herke van Hoof
2021 arXiv   pre-print
Given two sources of evidence about a latent variable, one can combine the information from both by multiplying the likelihoods of each piece of evidence.  ...  We study this problem in the setting with two conflicting reward functions learned from different sources.  ...  What should the agent believe about the true reward θ * ? A natural idea is to combine the information from these diverse sources. However, the sources may conflict.  ... 
arXiv:2103.12142v1 fatcat:jnqlwj5eo5ajtbwhusn3oh54gi

A Conceptual Framework for Externally-influenced Agents: An Assisted Reinforcement Learning Review [article]

Adam Bignold, Francisco Cruz, Matthew E. Taylor, Tim Brys, Richard Dazeley, Peter Vamplew, Cameron Foale
2020 arXiv   pre-print
These include heuristic reinforcement learning, interactive reinforcement learning, learning from demonstration, transfer learning, and learning from multiple sources, among others.  ...  The proposed taxonomy details the relationship between the external information source and the learner agent, highlighting the process of information decomposition, structure, retention, and how it can  ...  must choose whether (i) to maintain a separate model for each information source, (ii) to combine the information from all sources into a single model, or (iii) a combination of both.  ... 
arXiv:2007.01544v1 fatcat:iepfl62fyfhudghvjjqhuunjqq

This Week in The Journal

Teresa Esch
2019 Journal of Neuroscience  
The fact that reward-related responses appeared earlier in VP than in rmCD indicates that VP neurons must receive information about reward size from a different source.  ...  How information about incentive and drive is combined and used to motivate behavior remains unclear, however.  ...  The fact that reward-related responses appeared earlier in VP than in rmCD indicates that VP neurons must receive information about reward size from a different source.  ... 
doi:10.1523/jneurosci.twij.39.10.2019 fatcat:c7bxiacbtzgfrjf5bh2xshldlm

Combinatorial brain decoding of people's whereabouts during visuospatial navigation

Hans P. Op de Beeck
2013 Frontiers in Neuroscience  
Highly successful decoding of routes followed through the maze was possible by combining information about multiple aspects of navigation events across time and across multiple cortical regions.  ...  Here we present a new approach, combinatorial brain decoding, in which we decode complex behavior by combining the information which we can retrieve from the neural signals about the many different sub-processes  ...  ACKNOWLEDGMENTS This work was supported by grant G.0562.10 from the Fund for Scientific Research FWO-Flanders, and an interdisciplinary research grant from KU Leuven (IDO/06/004).  ... 
doi:10.3389/fnins.2013.00078 pmid:23730269 pmcid:PMC3657635 fatcat:sagey7uzknaitf7qb252t5h3hm

Reward associations and spatial probabilities produce additive effects on attentional selection

Beth A. Stankevich, Joy J. Geng
2014 Attention, Perception & Psychophysics  
We therefore pitted spatial probabilities against reward associations and found that the two sources of information had independent and additive effects.  ...  Critically, it has yet to be understood how multiple sources of selection history interact when presented simultaneously.  ...  For example, if all sources of information impact attentional control through a single mechanism, having multiple such sources should produce interactive effects on behavior.  ... 
doi:10.3758/s13414-014-0720-5 pmid:24944105 fatcat:3ehh7tnkzrglvjyficdqblvbui

Automatic weight learning for multiple data sources when learning from demonstration

B.D. Argall, B. Browning, M. Veloso
2009 2009 IEEE International Conference on Robotics and Automation  
Most Learning from Demonstration work to date considers data from a single teacher. In this paper, we consider the incorporation of demonstrations from multiple teachers.  ...  We introduce Demonstration Weight Learning (DWL) as a Learning from Demonstration algorithm that explicitly represents multiple data sources and learns to select between them, based on their observed reliability  ...  Our work is unique in combining the two LfD considerations of multiple data sources and reliability.  ... 
doi:10.1109/robot.2009.5152668 dblp:conf/icra/ArgallBV09 fatcat:gwanc62fsncazcdb7tkjou6pqq

FROMS: Feedback Routing for Optimizing Multiple Sinks in WSN with Reinforcement Learning

Anna Forster, Amy L. Murphy
2007 2007 3rd International Conference on Intelligent Sensors, Sensor Networks and Information  
Such a situation arises in WSNs with multiple, possibly mobile users collecting data from a monitored area.  ...  In this work, we describe how information local to each node can be shared without overhead as feedback to neighboring nodes, enabling efficient routing to multiple sinks.  ...  Our evaluation clearly shows that the additional expense of learning, combined with the negligible overhead to piggyback reward information significantly lowers routing cost.  ... 
doi:10.1109/issnip.2007.4496872 fatcat:x33fiz5lzrf65f6awxd6u5upwe

Continuous Control Monte Carlo Tree Search Informed by Multiple Experts

Joose Julius Rajamaki, Perttu Hamalainen
2018 IEEE Transactions on Visualization and Computer Graphics  
The tree search utilizes information from multiple sources including two machine learning models.  ...  In this paper we also present a new way to combine information from the various sources such that minimal amount of information is lost.  ...  We also propose a new way of combining information from multiple sources.  ... 
doi:10.1109/tvcg.2018.2849386 pmid:29994613 fatcat:q3z2gwrpjvbcvetxqjk4gpugwy

The integration of social influence and reward: Computational approaches and neural evidence

Damon Tomlin, Andrea Nedic, Deborah A. Prentice, Philip Holmes, Jonathan D. Cohen
2017 Cognitive, Affective, & Behavioral Neuroscience  
These findings add to our understanding of how the separate influences of reward from the environment and information derived from other social agents are combined to produce decisions.  ...  In addition, our findings suggest how these sources were processed and combined during decision-making.  ...  was combined with information derived from the participant's own experience.  ... 
doi:10.3758/s13415-017-0512-1 pmid:28540647 fatcat:sllfrsxffnhs3ltrrfm2lk26ra

Rewards and Challenges of eDNA Sequencing with Multiple Genetic Markers for Marine Observation Programs

Kathleen Pitz, Collin Closek, Anni Djurhuus, Reiko Michisaki, Kristine Walz, Alexandria Boehm, Mya Breitbart, Ryan Kelly, Francisco Chavez
2017 Biodiversity Information Science and Standards  
However, it is still unclear how best to compare and combine this information with morphological counts in order to inform policies and biodiversity metrics that are based on traditional sampling results  ...  However, it is still unclear how best to compare and combine this information with morphological counts in order to inform policies and biodiversity metrics that are based on traditional sampling results  ...  In order to combine data from multiple markers, species occupancy modeling was used to determine the probability that an OTU is truly present in a sample (as described in Kelly et al. 2017 and Lahoz-Monfort  ... 
doi:10.3897/tdwgproceedings.1.20548 fatcat:s5om6w6t3vh6velhxpbxjdegay

Exploiting Reinforcement Learning for Multiple Sink Routing in WSNs

A. Egorova-Forster, A.L. Murphy
2007 2007 IEEE Internatonal Conference on Mobile Adhoc and Sensor Systems  
To deal with the dynamics of the environment arising from mobility and failures, we choose a reinforcement learning approach where neighboring nodes exchange small amounts of information allowing them  ...  In this work, we focus on routing data to multiple, possibly mobile sinks.  ...  Our evaluation clearly shows that the additional expense of learning, combined with the negligible overhead to piggyback reward information significantly improves network lifetime.  ... 
doi:10.1109/mobhoc.2007.4428632 dblp:conf/mass/Egorova-ForsterM07 fatcat:3bw7frnkqbb3ppnaoamusrfuwu

Pediatric functional magnetic resonance neuroimaging: tactics for encouraging task compliance

Michael W Schlund, Michael F Cataldo, Greg J Siegle, Cecile D Ladouceur, Jennifer S Silk, Erika E Forbes, Ashley McFarland, Satish Iyengar, Ronald E Dahl, Neal D Ryan
2011 Behavioral and Brain Functions  
Discussion: We proposed that some forms of task noncompliance may emerge from less than optimal reward protocols.  ...  Essential features of the approach include a preference assessment for identifying multiple individualized rewards, increasing reinforcement rates during imaging by pairing tasks with chosen rewards and  ...  Accordingly, conventional reward-based protocols often employ multiple sources and different kinds of rewards to encourage motivation and task completion.  ... 
doi:10.1186/1744-9081-7-10 pmid:21548928 pmcid:PMC3113722 fatcat:3sjh76ed2zgxtihm6fz2wiw2dy

ShareBoost: Boosting for Multi-view Learning with Performance Guarantees [chapter]

Jing Peng, Costin Barbu, Guna Seetharaman, Wei Fan, Xian Wu, Kannappan Palaniappan
2011 Lecture Notes in Computer Science  
However, they lack the ability to mine most discriminant information sources (or data types) for making predictions.  ...  Algorithms combining multi-view information are known to exponentially quicken classification, and have been applied to many fields.  ...  To do so, we must first specify a reward function for each information source.  ... 
doi:10.1007/978-3-642-23783-6_38 fatcat:vhy37conhbg7pctrbycp4hbmqe

Dynamic routing of task-relevant signals for decision making in dorsolateral prefrontal cortex

Christopher H Donahue, Daeyeol Lee
2015 Nature Neuroscience  
Thus, multiple types of neural signals are flexibly routed in the DLPFC so as to favor actions that maximize reward. npg We thank M. Hammond and P. Kurnath for technical support, and Z.  ...  We analyzed the neuronal activity in the DLPFC of monkeys performing a probabilistic reversal task where information about the probability and magnitude of reward was provided by the target color and numerical  ...  Finally, we examined whether the animals combined probability and magnitude information additively or multiplicatively.  ... 
doi:10.1038/nn.3918 pmid:25581364 pmcid:PMC5452079 fatcat:oxjhy6zwxzfevkp2k3sf3hjt2y

Paradox Elimination in Dempster–Shafer Combination Rule with Novel Entropy Function: Application in Decision-Level Multi-Sensor Fusion

Md Nazmuzzaman Khan, Sohel Anwar
2019 Sensors  
As a result, DS-based information fusion is very popular in decision-making applications, but original DS theory produces counterintuitive results when combining highly conflicting evidences from multiple  ...  Multiple examples are presented to show that the proposed method is effective in handling conflicting information in spatial domain.  ...  Dempster-Shafer Rule of Combination The purpose of data fusion is to summarize and simplify information rationally, obtained from independent and multiple sources.  ... 
doi:10.3390/s19214810 fatcat:btpzgzjri5er7jgwmckflf7tja
« Previous Showing results 1 — 15 out of 250,070 results