1,720,995 research outputs found
Fact-checking via path embedding and aggregation
No full textKnowledge graphs (KGs) are a useful source of background knowledge to (dis)prove facts of the form (s, p, o). The goal of this paper is to present the Fact checking via path Embedding and Aggregation (FEA) system. FEA starts by carefully collecting the paths between s and o that are most semantically related to the domain of p. It learns vectorized path representations, aggregates them according to different strategies, and use them to finally (dis)prove a fact. Our experiments show that our hybrid solution brings benefits in terms of performance
Semantic Similarity Functions and Measures
No full textSimilarity is the process through which two or more objects are compared in order to identify to what extent they are alike. This is useful, for instance, when there is the need to identify common elements among a set of objects or when a new object has to be classified w.r.t. the others. Similarity is a very generic notion spanning among different disciplines from psychology to computer science. In this paper, the focus will be on computational methods that exploit ontologies or search engines as primary source of «background» knowledge. In particular, after surveying on the most popular similarity measures recently proposed, the Similarity Library will be presented, which has the aim to provide researchers and practitioners with a flexible tool encompassing several similarity measures both between words and sentences
An approach to Ontology Mapping based on the Lucene search engine library
No full textRecently ontology mapping has been identified as a key issue in semantic-based technologies. A mapping algorithm aims at finding correspondences, and matching, between URIrefs (classes, relationships, instances) of a source ontology and a target ontology by combining several matching components (matchers) each of which relies on one or more ontology features (linguistic, structural, etc.). In this paper we propose an approach to ontology mapping based on the Lucene search engine library. We exploit Lucene features to build an index from a source ontology in which Lucene documents, gathering different kinds of information (name, value, comment, label, etc.) about URIrefs, are stored. Therefore mappings are derived by using values of the URIrefs of the target ontology as search arguments against the index created from the source ontology. Experimental results show the suitability of this approach in terms of Precision, Recall, F-Measure and execution time, as compared to other four approaches. © 2007 IEEE
Ontology: Querying languages and development
No full textOntologies are a key support to provide a shared understanding of knowledge domains. Their flexibility and usefulness is witnessed in a variety of scenarios, from knowledge-based retrieval to projects like the Semantic Web. The possibility to express semantic information gave birth to a new class of query languages specifically designed. We will give an overview of one of such languages called SPARQL. SPARQL is the W3C standard to query data represented in RDF, another W3C standard. By using examples, we illustrate the structure of a SPARQL query. Then, we give an overview of popular ontology management tools
Querying graphs with preferences
No full textThis paper presents GuLP a graph query language that enables to declaratively express preferences. Preferences enable to order the answers to a query and can be stated in terms of nodes/edge attributes and complex paths. We present the formal syntax and semantics of GuLP and a polynomial time algorithm for evaluating GuLP expressions. We describe an implementation of GuLP in the GuLP-it system, which is available for download. We evaluate the GuLP-it system on real-world and synthetic data. Copyright 2013 ACM
CAKES: Cross-lingual wikipedia knowledge enrichment and summarization
No full textWikipedia is a huge source of multilingual knowledge curated by human contributors. Wiki articles are independently written in the various languages and may cover different perspectives about a given subject. The aim of this paper is to exploit Wikipedia multilingual information for knowledge enrichment and summarization. Investigating the link structure of a Wiki article in a source language and comparing it with the structure of articles about the same subject written in other languages gives insights about the body of knowledge shared among languages. This investigation is also useful to identify knowledge perspectives not covered in the source language but covered in other languages. We implemented these ideas in CAKES, which: i) exploits Wikipedia information on the fly without requiring any data preprocessing; ii) enables to specify the set of languages to be considered and; iii) ranks subjects interesting for a given article on the basis of their popularity among languages
From Node Embeddings to Triple Embeddings
No full textAn extended version of this paper has been published at the the 34th AAAI Conference on Artificial Intelligence (AAAI) with the title “Learning Triple Embeddings from Knowledge Graphs”. Graph embedding techniques allow to learn high-quality low-dimensional graph representations useful in various tasks, from node classification to clustering. Knowledge graphs are particular types of graphs characterized by several distinct types of nodes and edges. Existing knowledge graph embedding approaches have only focused on learning embeddings of nodes and predicates. However, the basic piece of information stored in knowledge graphs are triples and thus, an interesting problem is that of learning embeddings of triples as a whole. In this paper we report on Triple2Vec, a new technique to directly compute triple embeddings in knowledge graphs. Triple2Vec leverages the idea of line graph and extends it to the context of knowledge graphs. Embeddings are then generated by adopting the SkipGram model, where sentences are replaced with walks on a wighted version of the line graph
UFOme: A user friendly ontology mapping environment
No full textRecently the Ontology Mapping Problem (OMP) has been identified as a key factor towards the success of the Semantic Web and related applications. This problem arises since it is possible for different people to give, through ontologies, different conceptualizations of the same (or overlapping) knowledge domain. In order to tackle the OMP several algorithms have been designed. They aim at discovering correspondences (aka mappings) between ontology entities. However, these algorithms mostly suffer from the following shortcomings: (i) do not allow to quickly combine and/or compare different mapping strategies; (ii) do not offer support for evaluating mapping strategies in terms of quality of results and performance. In this paper we present a plugin-based system called UFOme along with its current implementation. We illustrate how it can be exploited to graphically design mapping tasks by connecting different types of modules. UFOme provides three categories of modules. The first one (i.e., visualization) allows to explore the ontologies to be mapped. The second one (i.e., matching) provides different types of individual matchers, exploited to discover mappings between ontologies, and a module for combining them. The third one (i.e., evaluation) enables to evaluate each module of the mapping task, a sub mapping task, or the mapping task in the whole w.r.t performance and quality of results
Ontology: Definition languages
No full textOntologies are artifacts used to model and represent in an explicit way knowledge related to a particular domain in terms of concepts, relations between concepts and axioms. In this article we provide an overview of ontology languages that have been defined under the umbrella of the W3C consortium. These language are characterized by different levels of expressiveness, starting from RDF, a simple language to express statements in the form of triples, to OWL, which enables very expressive forms of inference
Structural Characterization of Graph Navigational Languages
No full textGraph navigational languages define binary relations in terms of pair of nodes in a graph subject to the existence of a path satisfying a certain regular expression. The goal of this paper is to give a novel characterization of navigational languages in terms of the structure of the graph embracing the results of a query. We define novel graph-based query evaluation semantics and efficient algorithms able to represent and capture intermediate nodes/edges linking pairs of nodes in the answer. We enhance the language of Nested Regular Expressions (NREs) with our machineries, thus defining the language of Structural NREs (sNREs)
- …
