Author name cluster

Stijn Heymans

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

AAAI Conference 2014 Conference Paper

Large-Scale Analogical Reasoning

Vinay Chaudhri
Stijn Heymans
Adam Overholtzer
Aaron Spaulding
Michael Wessel

Cognitive simulation of analogical processing can be used to answer comparison questions such as: What are the similarities and/or differences between A and B, for concepts A and B in a knowledge base (KB). Previous attempts to use a general-purpose analogical reasoner to answer such questions revealed three major problems: (a) the system presented too much information in the answer, and the salient similarity or difference was not highlighted; (b) analogical inference found some incorrect differences; and (c) some expected similarities were not found. The cause of these problems was primarily a lack of a well-curated KB and, and secondarily, algorithmic deficiencies. In this paper, relying on a wellcurated biology KB, we present a specific implementation of comparison questions inspired by a general model of analogical reasoning. We present numerous examples of answers produced by the system and empirical data on answer quality to illustrate that we have addressed many of the problems of the previous system.

PDF Details

AAMAS Conference 2013 Conference Paper

"What If There Was No Oxygen? ": Responding to Hypothetical Questions in an Intelligent Tutoring Agent

Neil Yorke-Smith
Stijn Heymans
Vinay Chaudhri

Our aim is for intelligent tutoring agents to replace traditional and even online textbooks with personalized, adaptive, one-to-one instruction. We focus on science subjects, and describe an approach to answering hypothetical questions from the student, such as “Would cellular respiration continue in the absence of oxygen? ”

PDF

ECAI Conference 2010 Conference Paper

Tractable Reasoning with DL-Programs over Datalog-rewritable Description Logics

Stijn Heymans
Thomas Eiter
Guohui Xiao 0001

The deployment of KR formalisms to the Web has created the need for formalisms that combine heterogeneous knowledge bases. Nonmonotonic dl-programs provide a loose integration of Description Logic (DL) ontologies and Logic Programming (LP) rules with negation, where a rule engine can query an ontology with a native DL reasoner. However, even for tractable dl-programs, the overhead of an external DL reasoner might be considerable. To remedy this, we consider Datalog-rewritable DL ontologies, i. e. , ones that can be rewritten to Datalog programs, such that dl-programs can be reduced to Datalog ¬, i. e, Datalog with negation, under well-founded semantics. To illustrate this framework, we consider several Datalog-rewritable DLs. Besides fragments of the tractable OWL 2 Profiles, we also present [Lscr ][Dscr ][Lscr ] + as an interesting DL that is tractable while it has some expressive constructs. Our results enable the usage of DBLP technology to reason efficiently with dl-programs in presence of negation and recursion, as a basis for advanced applications.

Details

ECAI Conference 2006 Conference Paper

Approximating Extended Answer Sets

Davy Van Nieuwenborgh
Stijn Heymans
Dirk Vermeir

We present an approximation theory for the extended answer set semantics, using the concept of an approximation constraint. Intuitively, an approximation constraint, while satisfied by a “perfect” solution, may be left unsatisfied in an approximate extended answer set. Approximations improve as the number of unsatisfied constraints decreases. We show how the framework can also capture the classical answer set semantics, thus providing an approximative version of the latter.

Details

JELIA Conference 2004 Conference Paper

Hierarchical Decision Making by Autonomous Agents

Stijn Heymans
Davy Van Nieuwenborgh
Dirk Vermeir

Abstract Often, decision making involves autonomous agents that are structured in a complex hierarchy, representing e. g. authority. Typically the agents share the same body of knowledge, but each may have its own, possibly conflicting, preferences on the available information. We model the common knowledge base for such preference agents as a logic program under the extended answer set semantics, thus allowing for the defeat of rules to resolve conflicts. An agent can express its preferences on certain aspects of this information using a partial order relation on either literals or rules. Placing such agents in a hierarchy according to their position in the decision making process results in a system where agents cooperate to find solutions that are jointly preferred. We show that a hierarchy of agents with either preferences on rules or on literals can be transformed into an equivalent system with just one type of preferences. Regarding the expressiveness, the formalism essentially covers the polynomial hierarchy. E. g. the membership problem for a hierarchy of depth n is \(\sum{_{n+2}^P}\) -complete. We illustrate an application of the approach by showing how it can easily express a generalization of weak constraints, i. e. “desirable” constraints that do not need to be satisfied but where one tries to minimize their violation.

Details

LPAR Conference 2004 Conference Paper

Weighted Answer Sets and Applications in Intelligence Analysis

Davy Van Nieuwenborgh
Stijn Heymans
Dirk Vermeir

Abstract The extended answer set semantics for simple logic programs, i. e. programs with only classical negation, allows for the defeat of rules to resolve contradictions. In addition, a partial order relation on the program’s rules can be used to deduce a preference relation on its extended answer sets. In this paper, we propose a “quantitative” preference relation that associates a weight with each rule in a program. Intuitively, these weights define the “cost” of defeating a rule. An extended answer set is preferred if it minimizes the sum of the weights of its defeated rules. We characterize the expressiveness of the resulting semantics and show that it can capture negation as failure. Moreover the semantics can be conveniently extended to sequences of weight preferences, without increasing the expressiveness. We illustrate an application of the approach by showing how it can elegantly express subgraph isomorphic approximation problems, a concept often used in intelligence analysis to find specific regions of interest in a large graph of observed activities.

Details