Navegación

Búsqueda

Búsqueda avanzada

El autor Daniel Ayala ha publicado 2 artículo(s):

1 - AYNEC-DataGen: a tool for generating evaluation datasets for Knowledge Graphs completion

In the context of knowledge graphs, the task of completion of relations consists in adding missing triples to a knowledge graph, usually by classifying potential candidates as true of false. Creating an evaluation dataset for these techniques is not trivial, since there is a large amount of variables to consider which, if not taken into account, may cause misleading results. So far, there is not a well defined workflow that identifies the variation points when creating a dataset, and what are the possible strategies that can be followed in each step. Furthermore, there are no tools that help create such datasets in an easy way. To address this need, we have created AYNEC-DataGen, a customisable tool for the generation of datasets with multiple variation points related to the preprocessing of the original knowledge graph, the splitting of triples into training and testing sets, and the generation of negative examples. The output of our tool includes the evaluation dataset, an optional export in an open format for its visualisation, and additional files with metadata. Our tool is freely available online.

Autores: Daniel Ayala / Agustin Borrego / Inma Hernandez / Carlos R. Rivero / David Ruiz / 
Palabras Clave: Evaluation - Graph refinement - Knowledge Graph - Tool

2 - TAPON: a two-phase machine learning approach for semantic labelling

Through semantic labelling we enrich structured information from sources such as HTML pages, tables, or JSON files, with labels to integrate it into a local ontology. This process involves measuring some features of the information and then finding the classes that best describe it. The problem with current techniques is that they do not model relationships between classes. Their features fall short when some classes have very similar structures or textual formats. In order to deal with this problem, we have devised TAPON: a new semantic labelling technique that computes novel features that take into account the relationships. TAPON computes these features by means of a two-phase approach. In the first phase, we compute simple features and obtain a preliminary set of labels (hints). In the second phase, we inject our novel features and obtain a refined set of labels. Our experimental results show that our technique, thanks to our rich feature catalogue and novel modelling, achieves higher accuracy than other state-of-the-art techniques.

Autores: Daniel Ayala / Inma Hernandez / David Ruiz / Miguel Toro / 
Palabras Clave: Information Integration - Machine Learning - Semantic labelling