RDF Summarization
RDF Summarization
RDF Summarization
2. RDF definition
3. Advantages of RDF
5. RDF summarization
RDF definition
Advantages of
RDF
Problems in RDF RDF was original writeen by Tim Bray in 1998 an update by Dan Brickley in
representation
2001[15].
RDF
summarization
RDF summary
requerimients
The Resource Description Framework (RDF) is a lenguage for representing in-
Main Approaches for RDF formation about resources in the World Wide Web[15].
Summarization
Advantages of
RDF RDF descriptions, often contain redundancies, and could be generated differ-
summarization ently even when describing the same resources, which would have a negative
Our Proposal: impact on various RDF-based applications (e.g.,RDF storage, processing time,
RDF graph loading time, similarity measuring, mapping, alignment, and versioning)[13].
summarization
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF definition
RDF previous
Problems in RDF
representation
Our Proposal:
RDF graph IRI (Internationalized Resource Identifier) refers to a resource (the
summarization referent).
with RNN
RDF summarization with
RNN
Literal denotes resources which have an associated value for exam-
Saturation ple, an integer or string value.
Feature extraction
Recurrent Neural
Networks
Blank nodes are local identifiers which do not identify specific re-
References
sources.
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF definition
RDF previous
RDF definition
Advantages of
RDF
RDF data graph and RDF schema graph:
Problems in RDF
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
(a) RDF graph (b) RDF Shema (RDFS) graph
summarization
Our Proposal:
RDF graph
summarization
Figure 3.1: Example of RDF graph and RDFS graph
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF definition
RDF previous
RDF definition
RDFS Statement Triple Shorthand
Advantages of RDF Statement Triple Shorthand SubClass (s,rdf:subClassOf,o) (s,ăsc ,o)
RDF
Class Assertion (s,rdf:type,o) (s,t,o) Property assertion (s,rdf:subPropertyOf,o) (s,ăsp ,o)
Problems in RDF Property assertion (s,p,o) with p “ rdf:type (s,p,o) Domain typing (s,rdf:domain,o) (s,Ðâd ,o)
Range typing (s,rdf:range,o) (s,Ðâd ,o)
representation
RDF
summarization
Table 1: RDF & RDFS statements
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
summarization
Our Proposal:
RDF graph
summarization
with RNN
RDF summarization with
RNN (a) RDF graph (b) RDF Shema (RDFS) graph
Saturation
Feature extraction
Recurrent Neural
Networks Figure 4.1: Example of RDF graph and RDFS graph
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Advantages of RDF
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Problems in RDF representation
RDF previous
RDF definition
Advantages of
RDF
RDF 2 Minimum Constraints for RDF data make it irregular, difficult to comprehend
summarization and visualize, this can cause problems for information extraction, processing,
RDF summary
requerimients and analysis.
Main Approaches for RDF
Summarization
3 RDFs have been designed as a query standard based on explicit and implicit
Advantages of
RDF
data.
summarization
Our Proposal:
RDF graph
summarization
Many authors propose use RDF summarization
with RNN
RDF summarization with
approach to solve problem item 2 y 3
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF summarization
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
RDF RDF graph are often large and varied, produced in a variety of contexts such
summarization
RDF summary
as social networks, medical data, scientific data, etc. [9]. The large amount of
requerimients data contained in the RDF is often too expensive to perform queries to acquire
Main Approaches for RDF
Summarization information. The RDF summary refers to the process of extracting concise but
Advantages of
RDF significant summaries of RDF Knowledge Bases (KBs) that represent as close
summarization as possible the actual contents of the KB, both in terms of structure and data
Our Proposal: [17].
RDF graph
summarization
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF summary requerimients
RDF previous
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Main Approaches for RDF Summarization[17]
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
summarization
Our Proposal:
RDF graph
summarization
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
RDF definition
Advantages of
RDF
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Our Proposal: RDF graph summarization with
Recurrent Neural Network (RNN)
RDF previous
RDF definition In the paper: “Modeling Relational Data with Graph Convolutional Networks”[12]
Advantages of exist some impressions:
RDF
General Fourier transform scales poorly with size of data so we need relax-
Problems in RDF ations, normally use for image processing, where actual spatial convolutions
representation
are easy to compute.
RDF
summarization At all levels in this network, the filters are limited to 3x3 in size and are also
RDF summary
requerimients essentially fixed to be the same kernel across all layers and all units in entire
Main Approaches for RDF
Summarization
network
Advantages of
RDF In the paper “Convolutional Neural Networks on Graphs with Fast Localized
summarization Spectral Filtering”? ], the authors try to solve the above problems including
Our Proposal: higher order Chebyshev polynomials in the approximation.
RDF graph
summarization
with RNN
RDF summarization with
RNN
Saturation
For satisfy the stationarity, locality, compositionality assump-
Feature extraction
Recurrent Neural
tions, we propose use "Recurrent Neural Networks"
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
RDF summarization with Recurrent Neural
Network (RNN)
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
summarization
Our Proposal:
RDF graph
summarization
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks Figure 9.1: RDF graph summarization with RNN
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Saturation
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
summarization
Figure 9.3: Sample RDF
Our Proposal: entailment rules[5]
RDF graph
summarization Figure 9.2: Saturation of the graph union
with RNN
RDF summarization with
RNN
Saturation Other authors[14][12] assign labels to entity types, for example: “John Doe”,
Feature extraction
“London”, “England”, and “1967-01-10” can be mapped to “PERSON”, “CITY”,
Recurrent Neural
Networks “COUNTRY”, and “DATE”
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Feature extraction
RDF previous
RDF definition
Advantages of
RDF
Other technique for feacture
Problems in RDF extraction[1]
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF
Summarization
Advantages of
RDF
summarization
Our Proposal:
RDF graph
summarization
with RNN Figure 9.5: Example of aggregation using
RDF summarization with K-Snap
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
Figure 9.4: Feacture Extraction
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Recurrent Neural Networks
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation
RDF
summarization
RDF summary
requerimients
Main Approaches for RDF Figure 9.6: Recurrent Neural Networks Architecture
Summarization
Advantages of
RDF
summarization
Our Proposal:
RDF graph
summarization
with RNN
RDF summarization with
RNN Figure 9.7: Recurrent Neural Networks Architecture
Saturation
Feature extraction
Recurrent Neural
Networks Recurrent neural networks address the problem reasoning about previous events
References in the nodes to inform later ones. They are networks with loops in them, allowing
information to persist[6].
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
References I
RDF previous
RDF definition [1] Aufaure, A. and Louati, M. A. (2012). Graph Aggregation : Application to Social Networks Graph
Aggregation Algorithms.
Advantages of [2] Bursztyn, D., Goasdoué, F., and Manolescu, I. (2014). Optimizing Reformulation-based Query An-
RDF swering in RDF. 18th International Conference on Extending Database Technology (EDBT), pages
265–276.
Problems in RDF
representation [3] Cheng, G., Ji, F., Luo, S., Ge, W., and Qu, Y. (2012). BipRank: Ranking and summarizing RDF
vocabulary descriptions. Lecture Notes in Computer Science (including subseries Lecture Notes in
RDF Artificial Intelligence and Lecture Notes in Bioinformatics), 7185 LNCS:226–241.
summarization
RDF summary
[4] Doel, T., Shakir, D. I., Pratt, R., Aertsen, M., Moggridge, J., Bellon, E., David, A. L., Deprest, J.,
requerimients Vercauteren, T., and Ourselin, S. (2017). GIFT-Cloud: A data sharing and collaboration platform for
Main Approaches for RDF medical imaging research. Computer Methods and Programs in Biomedicine, 139:181–190.
Summarization
[5] Hassad, S. E. (2017). Learning Commonalities in RDF. pages 502–517.
Advantages of [6] Jozefowicz, R. and Com, I. G. (2015). An Empirical Exploration of Recurrent Network Architectures.
RDF 37.
summarization
[7] Khatchadourian, S. and Consens, M. P. (2010). ExpLOD : Summary-Based Exploration of Interlinking
Our Proposal: and RDF Usage in the Linked Open Data Cloud. pages 272–287.
RDF graph [8] Kushwaha, N., Singh, B., Mahule, R., and Vyas, O. P. (2015). Keyword prediction with ARM on biblio-
summarization graphic RDF data. Procedia Computer Science, 50:490–495.
with RNN
RDF summarization with [9] Manolescu, I. (2018). Structural Summarization of Semantic Graphs Ioana Manolescu To cite this
RNN version : HAL Id : hal-01808737 Structural Summarization of Semantic Graphs.
Saturation
[10] Ren, X., Curé, O., Naacke, H., and Xiao, G. (2018). BigSR: an empirical study of real-time expressive
Feature extraction
Recurrent Neural
RDF stream reasoning on modern Big Data platforms. pages 1–16.
Networks
[11] Ristoski, P. and Paulheim, H. (2016). RDF2Vec : RDF Graph Embeddings for Data Mining. page 30.
References [12] Schlichtkrull, M., Kipf, T. N., Bloem, P., Titov, I., and Welling, M. (2017). Modeling Relational Data
with Graph Convolutional Networks. (1).
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
References II
RDF previous
RDF definition
Advantages of
RDF
Problems in RDF
representation [13] Ticona, R., Tekli, J., Dongo, I., Guzman, R., and Chbeir, R. (2017). Toward RDF Normalization.
RDF [14] Trisedya, B. D., Qi, J., Zhang, R., and Wang, W. (2018). GTR-LSTM: A Triple Encoder for Sentence
summarization Generation from RDF Data. Proceedings of ACL, pages 1627–1637.
RDF summary
requerimients [15] w3c (2004). Resource description framework (rdf):concepts and abstract syntax. [Web; accedido el
Main Approaches for RDF 06-10-2018].
Summarization
[16] Zneika, M., Lucchese, C., Vodislav, D., and Kotzinos, D. (2016). Summarizing Linked Data RDF
Advantages of Graphs Using Approximate Graph Pattern Mining. Proc. 19th International Conference on Extending
RDF Database Technology, pages 684–685.
summarization
[17] Zneika, M., Vodislav, D., Kotzinos, D., Metrics, Q., Rdf, F., and Summarization, G. (2018). Quality
Our Proposal: Metrics For RDF Graph Summarization.
RDF graph
summarization
with RNN
RDF summarization with
RNN
Saturation
Feature extraction
Recurrent Neural
Networks
References
Luis Vives 1 , César Beltran 2 , Richard Chbeir 3 RDF SUMMARIZATION October 30, 2018
Thank You.