Filters








454 Hits in 4.7 sec

Planning with POMDPs using a compact, logic-based representation

Chenggang Wang, J. Schmolze
2005 17th IEEE International Conference on Tools with Artificial Intelligence (ICTAI'05)  
Representations built on logic allow for problems to be specified in a compact and transparent manner.  ...  In particular, we present a compact representation of POMDPs and a method to update beliefs after actions and observations.  ...  Moreover, we assume a finite domain. The representation scheme that follows is designed to allow compact representations using logic. .  ... 
doi:10.1109/ictai.2005.96 dblp:conf/ictai/WangS05 fatcat:o2jpm75u4jblvkbiaa2kyk54ee

Partially Observable Markov Decision Processes [chapter]

Nicole Bäuerle, Ulrich Rieder
2011 Universitext  
Representations built on logic allow for problems to be specified in a compact and transparent manner.  ...  In particular, we present a compact representation of POMDPs and a method to update beliefs after actions and observations.  ...  Moreover, we assume a finite domain. The representation scheme that follows is designed to allow compact representations using logic. .  ... 
doi:10.1007/978-3-642-18324-9_5 fatcat:uy4klthvpjcbnpsey4gwzn57se

Partially Observable Markov Decision Processes [chapter]

Alain Dutech, Bruno Scherrer
2013 Markov Decision Processes in Artificial Intelligence  
Representations built on logic allow for problems to be specified in a compact and transparent manner.  ...  In particular, we present a compact representation of POMDPs and a method to update beliefs after actions and observations.  ...  Moreover, we assume a finite domain. The representation scheme that follows is designed to allow compact representations using logic. .  ... 
doi:10.1002/9781118557426.ch7 fatcat:uhipce426zexbbfxp7nwvip6sq

Action representation and partially observable planning using epistemic logic

Andreas Herzig, Jérôme Lang, Pierre Marquis
2003 International Joint Conference on Artificial Intelligence  
We propose a purely logical framework for planning in partially observable environments. Knowledge states are expressed in a suitable fragment of the epistemic logic S5.  ...  We show how progression, regression and plan generation can be achieved in our framework.  ...  The representation of belief states in both approaches uses BDDs, which allow for a compact representation but not as space efficient as DAG-based propositional formulas.  ... 
dblp:conf/ijcai/HerzigLM03 fatcat:dmrzlznqo5djdhgrrl4snfcm24

Mixed Logical Inference and Probabilistic Planning for Robots in Unreliable Worlds

Shiqi Zhang, Mohan Sridharan, Jeremy L. Wyatt
2015 IEEE Transactions on robotics  
Nonmonotonic logical inference in ASP is used to generate a multinomial prior for probabilistic state estimation with the hierarchy of POMDPs.  ...  Robots need to represent and reason with incomplete domain knowledge, acquiring and using sensor inputs based on need and availability.  ...  This work was supported in part by the US Office of Naval Research Science of Autonomy award N00014-13-1-0766, and the European Commissionfunded Strands project FP7-IST-600623.  ... 
doi:10.1109/tro.2015.2422531 fatcat:lfz3lkoasvgmhaa74ty76zcyem

Towards Relational POMDPs for Adaptive Dialogue Management

Pierre Lison
2010 Annual Meeting of the Association for Computational Linguistics  
The tractability of the resulting POMDP can be preserved using a mechanism for dynamically constraining the action space based on prior knowledge over locally relevant dialogue structures.  ...  These constraints are encoded in a small set of general rules expressed as a Markov Logic network.  ...  POMDP-based dialogue management Dialogue management can be easily cast as a POMDP problem, with the state space being a compact representation of the interaction, the action space being a set of dialogue  ... 
dblp:conf/acl/Lison10 fatcat:xgcvq7uiurfunauyyktazu5abu

Knowledge-Based Policies for Qualitative Decentralized POMDPs

Abdallah Saffidine, François Schwarzentruber, Bruno Zanuttini
2018 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We propose and investigate a new representation for joint policies in QDec-POMDPs, which we call Multi-Agent Knowledge-Based Programs (MAKBPs), and which uses epistemic logic for compactly representing  ...  Qualitative Decentralized Partially Observable Markov Decision Problems (QDec-POMDPs) constitute a very general class of decision problems.  ...  As shown by Lang and Zanuttini (2012; for the single-agent case, KBPs offer compact representations of policies for planning problems.  ... 
doi:10.1609/aaai.v32i1.12085 fatcat:kmlxg2u4kjfmzpakg2oubvejrm

Combining Answer Set Programming and POMDPs for Knowledge Representation and Reasoning on Mobile Robots [article]

Shiqi Zhang, Mohan Sridharan
2013 arXiv   pre-print
Specifically, Answer Set Programming (ASP), a declarative programming paradigm, is combined with hierarchical partially observable Markov decision processes (POMDPs), using domain knowledge to revise probabilistic  ...  The architecture described in this paper combines declarative programming and probabilistic reasoning to address these challenges, enabling robots to: (a) represent and reason with incomplete domain knowledge  ...  logic formulas [12] , and Bayesian Logic (BLOG), which relaxes the unique name constraint of first-order probabilistic languages to provide a compact representation of probability distributions over outcomes  ... 
arXiv:1307.8084v1 fatcat:iyf4pftmwvdfnleyjt5ps6pqj4

Regular Decision Processes: A Model for Non-Markovian Domains

Ronen I. Brafman, Giuseppe De Giacomo
2019 Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence  
In RDPs, transition and reward functions are specified using formulas in linear dynamic logic over finite traces, a language with the expressive power of regular expressions.  ...  We introduce and study Regular Decision Processes (RDPs), a new, compact, factored model for domains with non-Markovian dynamics and rewards.  ...  Hence, we restrict attention to a class of NMDPs in which a finite, logic-based description is used to capture properties of traces.  ... 
doi:10.24963/ijcai.2019/766 dblp:conf/ijcai/BrafmanG19a fatcat:ljb4anx3bnas3cs7f4ywemvini

The Model-Based Approach to Autonomous Behavior: A Personal View

Hector Geffner
2010 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
is induced from experience via a learning algorithm, and the model-based approach, where the controller is derived from a model of the problem.  ...  In AI, three approaches have been used to address this problem: the programming-based approach, where the agent controller is given by the programmer, the learning-based approach, where the controller  ...  The languages used for representing conformant, contingent, MDP, and POMDP planning problems in compact form, are minor variations of the languages used for representing classical planning problems, in  ... 
doi:10.1609/aaai.v24i1.7765 fatcat:al7p7vu6uvhtpm5hqj2wa2kkki

Symbolic Dynamic Programming [chapter]

Scott Sanner, Kristian Kersting
2017 Encyclopedia of Machine Learning and Data Mining  
This is the first lifted relational POMDP solution that can optimally accommodate actions with a potentially infinite relational space of observation outcomes.  ...  Partially-observable Markov decision processes (POMDPs) provide a powerful model for sequential decision-making problems with partially-observed state and are known to have (approximately) optimal dynamic  ...  Case Representation of Reward, Value, & Probability: We use a tabular case notation along with its logical definition to allow first-order specifications of the rewards, probabilities, and values required  ... 
doi:10.1007/978-1-4899-7687-1_806 fatcat:vrzbo7ykcva2xbs3npbwd6qro4

Symbolic Dynamic Programming for First-order POMDPs

Scott Sanner, Kristian Kersting
2010 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
This is the first lifted relational POMDP solution that can optimally accommodate actions with a potentially infinite relational space of observation outcomes.  ...  Partially-observable Markov decision processes (POMDPs) provide a powerful model for sequential decision-making problems with partially-observed state and are known to have (approximately) optimal dynamic  ...  Case Representation of Reward, Value, & Probability: We use a tabular case notation along with its logical definition to allow first-order specifications of the rewards, probabilities, and values required  ... 
doi:10.1609/aaai.v24i1.7747 fatcat:kjbgf4upibbcjen32qbmu4gwoy

A Decentralized Partially Observable Decision Model for Recognizing the Multiagent Goal in Simulation Systems

Shiguang Yue, Kristina Yordanova, Frank Krüger, Thomas Kirste, Yabing Zha
2016 Discrete Dynamics in Nature and Society  
In this compact way, we only focus on the distribution of joint policies.  ...  In the experiments, a new scenario is designed based on the standard predator-prey problem: we increase the number of preys, and our aim is to recognize the real target of predators.  ...  Additionally, the increasing interests/number of works of planning theory based on Dec-POMDP can provide us with a large number of planners [24, 25] .  ... 
doi:10.1155/2016/5323121 fatcat:3nayudu2kra7lgy3jor2ir5ife

Automated Planning and Model Checking (Dagstuhl Seminar 14482)

Alessandro Cimatti, Stefan Edelkamp, Maria Fox, Daniele Magazzeni, Erion Plaku, Marc Herbstritt
2015 Dagstuhl Reports  
In artificial intelligence, automated planning deals with the automatic generation of plans for achieving a goal.  ...  There has been a lot of work on the exchanges between the areas of automated planning and model checking, based on the observation that a model-checking problem can be cast as a planning problem and vice-versa  ...  Sampling-based tree search with discrete abstractions for motion planning with dynamics and temporal logic.  ... 
doi:10.4230/dagrep.4.11.227 dblp:journals/dagstuhl-reports/CimattiEFMP14 fatcat:ukc75llezvg2jevqaytqouvzje

Solving Relational and First-Order Logical Markov Decision Processes: A Survey [chapter]

Martijn van Otterlo
2012 Adaptation, Learning, and Optimization  
We discuss model-freeboth value-based and policy-based -and model-based dynamic programming techniques.  ...  Such relational worlds can be found everywhere in planning domains, games, real-world indoor scenes and many more.  ...  left) a Blocks World planning problem and an optimal plan, (right) a logical Q-tree.  ... 
doi:10.1007/978-3-642-27645-3_8 fatcat:6yt5v5gbj5gsjab74lu6jypcxu
« Previous Showing results 1 — 15 out of 454 results