Dr. Jeff Z. Pan gave a tutorial in JIST2014 conference on Constructing and Understanding Knowledge Graph.
Abstract: The benefits and potentials of Linked Data (LD) have been utilised and demonstrated by numerous applications from academic, industry and public sectors. This explains the recent vast increase of LD not only in data volume but also in number of datasets and related domains. However, consuming a linked dataset requires technical background of Semantic Web (SW) techniques and the knowledge of the dataset. Direct use of such valuable knowledge space is very time consuming and is still a privilege of SW “geeks”. In this tutorial, we propose the vision of converting LD into knowledge graphs which are not only capable to enhance accessibilities in LD consumption but also enable LD directly usable to end users. Specifically, this tutorial consists of two parts. The first part will introduce the overview, applications and research challenges to create knowledge graphs on top of LD. The second part will focus on specific techniques for knowledge graph, including knowledge graph construction and knowledge graph understanding.
Speaker: Prof Thomas Eiter (TU Wien)
Title: Towards a Logic-based Framework for Analyzing Stream Reasoning
2-3pm, 29th Oct. 2014
Abstract: The rise of smart applications has drawn interest to logical reasoning over data streams. Recently, different query languages and stream processing / reasoning engines were proposed in different communities. However, due to a lack of theoretical foundations, the expressivity and semantics of diverse approaches was given only informally. Towards clear specifications and means for analytic study, a formal framework is desired that allows to characterize their semantics in precise terms. Inspired by this, we present ideas on a logic-based such framework, which features rules and window operators that provide a flexible mechanism to represent views on streaming data. We briefly discuss some complexity issues and relationships to stream query / reasoning languages, in particular capturing of the Continuous Query Language (CQL). This work is part of an ongoing project funded by the Austrian Science Fund.
Bio: Thomas Eiter is a professor at TU Wien since 1998, where he heads the Knowledge Base Systems Group (KBS) and the Institute of Information Systems. Among his current research interests are knowledge representation and reasoning, computational logic, and declarative problem solving. He co-chaired various meetings, most recently KR 2014 and the Vienna Summer of Logic, the largest event in the history of logic. Eiter is an ECCAI Fellow and Corresponding Member of the Austrian Academy of Sciences, and current president of KR, Inc.
In ISWC2014 conference, Jeff Z. Pan gave a tutorial on Large Scale Reasoning Over Semantic Data.
Abstract: The tutorial aims to provide an overview of the approaches used for large scale reasoning over semantic data, the systems developed as well as the lessons learned while developing them. We will discuss some applications which require scalable reasoning solutions. Questions such as what makes distributed/parallel reasoning hard would also be covered during the tutorial. Directions for future research work would be discussed.
Dr. Jeff Z. Pan gave a talk to K-Drive colleagues in IBM Italy. The topic is “Towards ontology based hypothesis generation and validation”, which is related to WP8 and WP9 in K-Drive project.
In last week’s two-day K-Drive workshop, our Marie Curie Fellow, Alessandro Faraotti from IBM Italy, gave a talk about Watson Technology including DeepQA and evidence-based decision making. The detail information of this talk is as follows.
Title: An introduction to Cognitive Computing and Watson
Abstract: The presentation introduces the notion of Cognitive Computing, by specifying what Cognitive Computing is and why it is a step forward in computer science. In particular, it presents the Watson experience in playing the Jeopardy game, the underling Watson architecture and details of all key components.
Alessandro Faraotti and Marco Monti start their secondments from 22.09.2014. In these secondments, they will mainly work on work packages 5 and 6. The topics are dynamic query generation and stream data processing. Furthermore, they will also give presentations about IBM Watson technologies.
Speaker: Olivier Curé (University of Paris-Est) <http://igm.univ-mlv.fr/%7Eocure/LIGM_LIKE/>
Title: WaterFowl: a Compact, Self-indexed RDF Store based on Succinct Data Structures
Time: 24 September 2014, 14:00 – 15:00
Location: Meston 2 (ground floor)
Abstract:This talk will start with an introduction of the main strategies for storing and indexing RDF data sets. This will consider solutions based on a native RDF approach but also approaches using a relational or NoSQL storage backend. Then, I will present the main features of an on-going work that aims to distribute highly compressed structures adapted for the storage and querying of RDF triples. The compactness of the represented data is supported by an architecture based on Succinct Data Structures (SDS) which enables to store large datasets in main memory. A special form of entity encoding enables inferences in the RDFS entailment regime.
Bio:Olivier Curé is an Associate Professor at the University of Paris-Est Marne-la-Vallée. His research focuses primarily on Knowledge Representation, Reasoning and Database systems. He is particularly interested on bridging these different fields of Computer Science. Some of the results he has obtained have been applied to the medical domain, more precisely in self-medication. During the 2014-2015 year, he will be on a leave of absence and will be a researcher in the LIP6 Database group at the Paris VI Pierre and Marie Curie University.
Dr. José Manuel Gómez Pérez, one of our Marie Currier Fellows of K-Drive project, will give a talk in University of Aberdeen.
Title: When History Matters – Assessing Reliability for the Reuse of Scientific Workflows
Abstract: Scientific workflows play an important role in computational research, as the essential artifacts for communicating the methods used to produce the research findings. We are witnessing a growing number of efforts of treating workflows as first-class artifacts for sharing and exchanging scientific knowledge, either as part of scholarly articles or as stand-alone objects. However, workflows are not born to be reliable, which can seriously damage their reusability and trustworthiness as knowledge exchange instruments. Scientific workflows are commonly subject to decaying, which consequently undermines their reliability over their lifetime. The reliability of workflows can be notably improved by advocating scientists to preserve a minimal set of information that is essential to assist the interpretations of these workflows and hence improve their potential for reproducibility and reusability. In this talk we show how, by measuring and monitoring the completeness and stability of scientific workflows over time we are able to provide scientists with a measure of their reliability, supporting the reuse of trustworthy scientific knowledge.
Venue: Meston 2
Time:14:00-15:00, 29 Sep. 2014
We are pleased to announce that we will have a two-day K-Drive workshop (25-26th of September, 2014) in Aberdeen. The venue is Meeting Room 3, Sir Duncan Rice Library, University of Aberdeen.
Marie Curie Fellow, Yuan Ren from University of Aberdeen, gave a talk in IBM Italy on 8th of July, 2014. The detail was as follows.
Title: Ontology Stream Reasoning with Combined Forward and Backward Chaining Completion
Abstract: Due to the dynamic nature of knowledge and data in semantic applications, ontology stream reasoning technologies are essential for ontology management systems. Nowadays, many proposed stream reasoning solutions and implemented systems apply forward chaining completion algorithms to handle the removal and addition of axioms. This talk introduces a novel approach to ontology stream reasoning that combines forward and backward chaining completion. Compared to existing works, this approach can be applied with or without bookkeeping, does not affect parallelisation or tractability, and reduces the effort for re-deriving the deleted results both theoretically and empirically.