Learning probabilistic relational planning rules
Published 3 June 2004
Hanna Pasula, Luke Zettlemoyer, Leslie Pack Kaelbling
Citations50
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
Abstract
To learn to behave in highly complex domains, agents must represent and learn compact models of the world dynamics. In this paper, we present an algorithm for learning probabilistic STRIPS-like planning operators from examples. We demonstrate the effective learning of rule-based operators for a wide range of traditional planning domains.
Keywords
Computer Science
Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference
16,927 Citations1988Judea Pearl
The author provides a coherent explication of probability as a language for reasoning with partial belief and offers a unifying perspective on other AI approaches to uncertainty, such as the Dempster-Shafer formalism, truth maintenance systems, and nonmonotonic logic.
Artificial IntelligenceIntelligence without representation
4,625 Citations1991Rodney A. Brooks
The fundamental decomposition of the intelligent system is not into independent information processing units which must interface with each other via representations, but into independent and parallel activity producers which all interface directly to the world through perception and action.
STRIPS: a new approach to the application of theorem proving to problem solving
1,479 Citations1971Richard Fikes, Nils J. Nilsson
Inductive Logic Programming: Techniques and Applications
767 Citations1993Nada Lavrač, Sašo Džeroski
Applications of inductive logic programming: learning rules for early diagnosis of rheumatic diseases finite element mesh design an overview of selected ILP applications.
Principles of Knowledge Representation and ReasoningUCPOP: a sound, complete, partial order planner for ADL
684 Citations1992J. Scott Penberthy, Daniel S. Weld
It is proved ucpop is both sound and complete for this representation and a practical implementation that succeeds on all of Pednault's and McDermott’s examples, including the infa-mous "Yale Stacking Problem" [McDermott 1991].
arXiv (Cornell University)Context-Specific Independence in Bayesian Networks
553 Citations2013Craig Boutilier, Nir Friedman +2 more
Artificial IntelligenceStochastic dynamic programming with factored representations
365 Citations2000Craig Boutilier, Richard Dearden +1 more
This work uses dynamic Bayesian networks (with decision trees representing the local families of conditional probability distributions) to represent stochastic actions in an MDP, together with a decision-tree representation of rewards, and develops versions of standard dynamic programming algorithms that directly manipulate decision-Tree representations of policies and value functions.
arXiv (Cornell University)A Bayesian Approach to Learning Bayesian Networks with Local Structure
326 Citations2013David Maxwell Chickering, David Heckerman +1 more
Symbolic dynamic programming for first-order MDPs
211 Citations2001Craig Boutilier, Ray Reiter +1 more
This technique uses an MDP whose dynamics is represented in a variant of the situation calculus allowing for stochastic actions and produces a logical description of the optimal value function and policy by constructing a set of first-order formulae that minimally partition state space according to distinctions made by the valuefunction and policy.
Probabilistic planning with information gathering and contingent execution
193 Citations1994Denise L. Draper, Steve Hanks +1 more
This work extends the BURIDAN probabilistic planning algorithm, adapting the action representation to model the behavior of imperfect sensors, and combines it with a framework for contingent action that extends the CNLP algorithm for conditional execution.
Elsevier eBooksLearning by Experimentation: Incremental Refinement of Incomplete Planning Domains
132 Citations1994Yolanda Gil
The paper shows that, contrary to popular belief, missing information is not necessarily associated with and selective interaction with the environment that pinpoints the type of fault in the domain knowledge that causes any unexpected behavior of the environment, and resorts to experimentation when additional information is needed to correct fault.
Lecture notes in computer scienceProbabilistic Planning in the Graphplan Framework
129 Citations2000Avrim Blum, John Langford
An MDP-style framework in which the state of the world is known but actions are probabilistic, and the objective is to produce a finite horizon contingent plan with highest probability of success within the horizon is considered.
Learning statistical models from relational data
121 Citations2011Lise Getoor, Lilyana Mihalkova
This survey will provide an introduction to SRL, describe some of the distinguishing characteristics of SRL systems, including relational feature construction and collective classification, and discuss connections between SRL methods and existing database research such as probabilistic databases.
Probabilistic reasoning for complex systems
107 Citations1999Daphne Koller, Avrom Jacob Pfeffer
This thesis presents a powerful object-based representation language, that integrates logical and probabilistic representations, and presents inference algorithms for the language that exploit the structure that can be expressed in it—not only the conditional independence structure normally exploited by BN algorithms, but also encapsulation, reuse of computation and symmetry resulting from the object- based representation.
Learning action models for reactive autonomous agents
74 Citations1996Scott Benson
This thesis describes TRAIL (Teleo-Reactive Agent with Inductive Learning), an integrated agent architecture which learns models of actions based on experiences in the environment which are then used to create plans that combine both goal-directed and reactive behaviors.
Rule creation and rule learning through environmental exploration
69 Citations1989Wei‐Min Shen, Herbert A. Simon
This paper reports an approach in which exploration, rule creation and rule learning are coordinated in a single framework, which creates STRIPS-Iike rules by noticing the changes in the environment when actions are taken, and later refines the rules by explaining the failures of their predictions.
Dynamic probabilistic relational models
63 Citations2003Sumit Sanghai, Pedro Domingos +1 more
This paper successfully applies dynamic probabilistic relational models (DPRMs) to execution monitoring and fault diagnosis of an assembly plan, in which a complex product is gradually constructed from subparts.
Probabilistic partial evaluation: exploiting rule structure in probabilistic inference
47 Citations1997David Poole
This paper presents a representation of the conditional probability tables of a variable given its parents in terms of parent contexts and provides an algorithm that exploits this compactness.
Searching for planning operators with context-dependent and probabilistic effects
36 Citations1996Tim Oates, Paul R. Cohen
Empirical results show that the algorithm successfully finds operators that capture the true structure of an agent's interactions with its environment, and avoids spurious associations between actions and exogenous events.
Elsevier eBooksEfficient Domain-Independent Experimentation
18 Citations1993Yolanda Gil
Domain-independent heuristics that compare possible hypotheses and choose the ones most likely to cause the failure are described that are implemented in EXPO, a system that uses PRODIGY as a baseline planner and improves its domain knowledge in several domains.
Learning and Planning with Probabilistic Relational Rules
5 Citations2004Hanna Pasula, Luke Zettlemoyer +2 more
A three-dimensional blocks-world simulation built with the OpenDynamics toolkit that represents world action dynamics using probabilistic planning rules to take advantage of the inherent structure found in many uncertain, complex environments.