43 search results - page 3 / 9 » Efficiently Constructing Relational Features from Background... |
ML
2006
ACM
13 years 7 months ago
2006
ACM
We propose a simple approach to combining first-order logic and probabilistic graphical models in a single representation. A Markov logic network (MLN) is a first-order knowledge b...
ACL
2006
13 years 8 months ago
2006
We present a novel hybrid approach for Word Sense Disambiguation (WSD) which makes use of a relational formalism to represent instances and background knowledge. It is built using...
KES
2007
Springer
2007
Springer
Inductive Concept Retrieval and Query Answering with Semantic Knowledge Bases Through Kernel Methods
14 years 1 months ago
This work deals with the application of kernel methods to structured relational settings such as semantic knowledge bases expressed in Description Logics. Our method integrates a n...
ICML
2005
IEEE
14 years 8 months ago
2005
IEEE
Markov logic networks (MLNs) combine logic and probability by attaching weights to first-order clauses, and viewing these as templates for features of Markov networks. In this pap...
NLP
2000
13 years 11 months ago
2000
In this article we compare the performance of various machine learning algorithms on the task of constructing word-sense disambiguation rules from data. The distinguishing characte...