Dr. Jeff Pan gave a seminar at IBM Italy

Dr. Jeff Pan gave a seminar at IBM Italy on 19th, September, 2013. His talk is titled:

It noticing so, Optimal and “Good Enough” Approximations for Semantic Reasoning.

Abstract: The talk will illustrate the challenges of reasoning with semantic data, discuss different reasoning strategies, and particularly focus on the approximation techniques. Two approximation technologies, namely the optimal semantic approximation and the “good enough” syntactic approximation, and their applications in the context of the K-Drive project, will be introduced.

Yuan Ren gave a seminar at IBM Italy

Yuan Ren gave a seminar at IBM Italy on 13th September, 2013. His talk is titled: Ontological Representation of Treatment Programs and Deviations.

Abstract: The talk will report the analysis results of the IBM Italy use case in the K-Drive project and its connections to knowledge-based technologies. Particularly, an ontological solution based on semantic query answering has been proposed for the medical treatment program deviation analysis use case. The construction of different knowledge assets in this use case, particularly the development of different queries for detecting deviations, will be illustrated.

Dr. Jeff Pan is in IBM Italy for his secondment

Dr. Jeff Z. Pan, Reader at the Department of Computing Science in the University of Aberdeen, is in IBM Italy for his secondment now. He will visit Rome for a week and then the Advanced Center of Research in Trento for another week. During his stay, Dr. Pan will work with researchers in IBM Italy and other secondee from University of Aberdeen on WP3, WP5 and WP6. He will also discuss other collaboration opportunities with IBM Italy.

Yuan Ren is in IBM Italy for his secondment

Yuan Ren from Aberdeen is now in the IBM Italy research center in Trento for his secondment of 8 weeks, 29th July – 20th September.

During his secondment at IBM Italy, Yuan will collaborate with the local researchers at IBM to:

  • Analyse the data used in their use case;
  • Analyse the requirements to K-Drive solutions, especially query generation and stream query answering technologies;
  • Extend the deliverables D2.1, 2.2, 2.3, 3.1 based on the above analysis;
  • Develop query generation technologies and draft deliverable D3.2;
  • Kick-off the work on WP5 and WP6;

Yuan as an experienced researcher will also supervise early-stage researcher secondees, if any, to work better on their secondments.

New Paper Published and Presented on ISWC 2012

A new paper partially funded by K-Drive – “RDFS Reasoning on Massively Parallel Hardware” is published and presented on this year’s International Semantic Web Conference.

Abstract. Recent developments in hardware have shown an increase in parallelism as opposed to clock rates. In order to fully exploit these new avenues of performance improvement, computationally expensive workloads have to be expressed in a way that allows for fine-grained parallelism. In this paper, we address the problem of describing RDFS entailment in such a way. Different from previous work on parallel RDFS reasoning, we assume a shared memory architecture. We analyze the problem of duplicates that naturally occur in RDFS reasoning and develop strategies towards its mitigation, exploiting all levels of our architecture. We implement and evaluate our approach on two real-world datasets and study its performance characteristics on different levels of parallelization. We conclude that RDFS entailment lends itself well to parallelization but can benefit even more from careful optimizations that take into account intricacies of modern parallel hardware.

You can view a draft of this paper on our Publications page.

Visiting Student arrives in the Aberdeen University

Today, Man Zhu, a PhD student from Southeast University China, starts a 6-month visit at the Aberdeen University. During her stay in Aberdeen, she will contribute to the research of K-Drive. Man Zhu is mainly interested in statistical relational learning and semantic web data investigation. Her paper “Ontology Learning from Noisy Linked Data” wins the best doctoral consortium paper in ISWC 2011. As part of the knowledge transfer activities in K-Drive, she will be involved in the development of novel data summarisation and query generation technologies.

New Secondment Starts!

Today, K-Drive researcher Dr. José Manuel Gómez-Pérez (from the iSOCO) starts his secondment at the University of Aberdeen. This secondment is related to WP1: Case Study on Exploitation of Public Semantic Data. During his secondment at Aberdeen, José will collaborate with the local researchers to finalise the analysis of some enterprise and public datasets used in the K-Drive project. This will help build up the use case regarding public semantic data and also contribute to the definition of requirements for technologies developed in later WPs.

New Paper Published and Presented on AAAI 2012

A new paper partially funded by K-Drive – “Querying Linked Ontological Data through Distributed Summarization” is published and presented on this year’s AAAI conference.

Abstract. As the semantic web expands, ontological data becomes distributed over a large network of data sources on the Web. Consequently, evaluating queries that aim to tap into this distributed semantic database necessitates the ability to consult multiple data sources eciently. In this paper, we propose methods and heuristics to efficiently query distributed ontological data based on a series of properties of summarized data. In our approach, each source summarizes its data as another RDF graph, and relevant section of these summaries are merged and analyzed at query evaluation time. We show how the analysis of these summaries enables more efficient source selection, query pruning and transformation of expensive distributed joins into local joins

You can view a draft of this paper on our Publications page.