Hide menu

Seminars on Semantic Technologies

*** NOTE: this seminar series has been replaced by the Semantic Web seminars presented here ****

After being dormant since autumn 2013, due to my parental leave, we are now reviving the seminar series again during autumn 2014! Do you have a presentation that you think would interest others at IDA, and which falls into the (broad) area of "semantic technologies" - send an e-mail to eva.blomqvist((at))liu.se

This page lists the seminars on Semantic Technologies held at IDA and contains links to slides etc. It also contains the schedule for the seminar occasions planned in the future.

The seminars usually consist of an invited talk, and an "open floor" discussion. In the first part a speaker presents some research related to semantic technologies for about 30-40 minutes, whereafter we reserve at least 15 minutes for questions and discussion.

Join the mailing-list and discuss semantic technologies outside the seminars! Subscribe to semtech __at__ ida.liu.se

Upcoming Seminar:

November 20th 2014, 15:15 in Alan Turing
The theme of the seminar will be on semantic stream processing.

Talk 1 (approx. 30 min including questions):

Presenters: Robin Keskisärkkä

Title: Semantic Complex Event Processing

Abstract:The amount of information available as online streams is growing rapidly, ranging from social media and RSS feeds to surveillance and sensor streams. Recently, the Semantic Web community has attempted to bring streams to a semantic level, but traditional methods assume data to be more or less static. To tackle this problem a number of RDF Stream Processing (RSP) engines have been developed, aimed specifically at dealing with Linked Stream Data. In this talk, we will discuss semantically enabled event processing in the context of Semantic Web technologies. It should be noted that an event in this context should be viewed as a representation of anything that happens or is contemplated as happening, e.g. a temperature reading from a sensor or Twitter message. Furthermore, we will discuss how complex events (e.g. aggregated and layered events) can be generated from event streams to support detection of high-level events.

Talk 2 (approx. 30 min including questions):

Presenters: Daniel de Leng

Title: Semantic Information Integration in a Stream Reasoning Framework

Abstract: Building complex systems such as autonomous robots usually require the integration of a wide variety of components including high-level reasoning functionalities. One important challenge is integrating the information in a system by setting up the data flow between components so that each component gets the information it needs and the system generates the desired output. In this talk, we discuss semantic matching as an approach to adaptive on-demand semantic information integration based on semantic web technologies. We take two important stand-points. First, we consider streams of information, to handle the fact that information often becomes continually and incrementally available. Second, we explicitly represent the semantics of the components and the information that can be provided by them in an ontology. Based on the ontology semantic matching automatically sets up the stream processing needed to generate the streams of information requested. Since the ontology represents both the system's information about the world and its internal stream processing many other powerful forms of introspective reasoning are also made possible.

Planned Seminars (Autumn 2014):

December 2014, date and time TBD
Presenter: TBD

Past Seminars:

October 10th 2014, 13:15 in Alan Turing

Talk 1 (approx. 30 min including questions):
Presenters: Gyula Fóra and Márton Balassi

Title: Apache Flink and its stream processing engine

Abstract:Apache Flink is a platform for efficient, distributed, general-purpose data processing. Flink features powerful programming abstractions in Java and Scala, a high-performance runtime, and automatic program optimization. It has native support for iterations, incremental iterations, and programs consisting of large DAGs of operations. Flink also provides APIs for distributed graph and stream processing. Márton Balassi and Gyula Fóra are committers of Flink. Their main contribution is the Flink Streaming component, which is highlighted by the talk. For further information please refer to http://flink.incubator.apache.org/.

Talk 2 (approx. 30 min including questions):
Presenters: Rita Kovordanyi

Title: Event driven processing - experiences with Esper

Abstract: In event driven processing, the control flow (execution) of a program is driven by the intermittent arrival of data (think listeners in java). Esper is a java-based scripting language optimized to handle massive streams of events. The event stream can be filtered, aggregated, and reasoned with in real-time using SQL-like expressions. I will briefly describe my experiences with using Esper in a EU FP7 project for simulator-based training of cross-organizational crisis responses, and conclude with highlighting possible future extensions of Esper, including fuzzy matching and reasoning capabilities.

April 16th 2013, 10:00 - note the time! (Alan Turing)

Presenters: Bosse Andersson (Semator AB) and Marie Gustafsson Friberger (Malmö högskola)
Tentative title: Introduktion till länkade data - (Note! Seminar will be in Swedish!)
The seminar is the start of a meetup on Swedish linked data, you are welcome to also join the rest of the day: http://ldsv2013.eventbrite.com/ (Note! the language will be Swedish)

March 5th 2013, 13:15 (Alan Turing)

Presenter: JD Bothma, Uppsala University
Tentative title: Ontology Learning from Swedish texts

October 31st 2012, 15:30 (John von Neumann, IDA)

Presenter: Valentina Presutti, Semantic Technologies Lab, CNR, Rome, Italy

Title: Making sense of Web content with Knowledge Patterns

Abstract:One of the characteristics of the Web is the variety of its contents. They differ by format, conceptual model, semantics, by evolution speed, etc. Such variety is a strength as well as a weakness: on one hand, all such content sources make the Web the greatest source of open information ever. On the other hand, the automatic exploitation of such information in specific contexts is most of the time very hard, and relies on significant cognitive effort from users. We claim that knowledge patterns (KP) can be used as a "passe-partout" for interpreting and aggregating knowledge, in other words for making sense of Web contents. In this presentation, I will introduce the concept of KP, and show results achieved so far through experiments as well as prototypes developed in the last two years that deal with KP extraction from Wikipedia, KP-driven exploratory search, KP-based machine reading, etc.

Presentation slides.

October 15th 2012, 15:15 (Alan Turing, IDA)

Presenter: Robin Keskisärkkä

Title: Semantic Stream Reasoning for RDF data - State of the art and tools


Presentation slides (to appear).

May 15th 2012, 13:15 (Allen Newell, IDA)

Presenter: Fabio Ciravegna, University of Sheffield (UK)

Title: Situation awareness via social streams

Abstract: In the recent years we have witnessed a growing interest in using social streams (e.g. Twitter Feeds) for situation awareness during emergencies. Social streams reflect what citizens see or think. People are everywhere at any time, often where emergency responders cannot be. The literature shows promising results in efficiently and effectively detecting earthquakes [Sakaki 2010], enabling organization of citizens and communities in case of fires [Palen 2006], floods [Palen 2010], and social unrest. In this talk I will present and discuss our experience in the analysis of social streams for situation awareness. Our main aim is to provide means to identifying events (what), their timeline (when), their participants (who) and the locations involved (where), so to empower organisational intelligence, i.e. the ability to understand and respond to emergencies. Research areas include - among others - human computer interaction, high-speed information extraction from text and semantic technologies. The scientific challenge in analysing social streams is that data and information are: (i) high in volume, and constantly increasing, (ii) often duplicated, incomplete, imprecise and incorrect; (iii) written in informal style (i.e. short, unedited and conversational); and (iv) generally concerning the short-term zeitgeist. These characteristics make analysis very hard, especially when considering that major requirements of emergency responders are that (i) documents must be processed in real-time and (ii) the relevant information may be in the long-tail of the distribution, i.e. it may be mentioned very infrequently. Veracity of information and reliability of sources are other important issues. Social streams provide a fertile environment for the growing number of trolls, i.e. people who wilfully diffuse wrong information, often to hamper the emergency response itself. I will present: · The scientific challenges and the way we are addressing them · An experience in monitoring an emergency exercise in preparation for the Olympics · An experience in releasing a system to emergency responders in two cities during two large events

Speaker bio: Prof. Dr. Fabio Ciravegna is Director of Research and Innovation in the Digital World at the University of Sheffield. He is also Professor of Language and Knowledge Technologies and Head of OAK Group in the Department of Computer Science. His research field concerns Human Language Technologies, Semantic Web Technologies, with focus on their use for Knowledge Management. Since 2006, he has been Director of the European Integrated Project IST X-Media on large-scale knowledge management across media and principal investigator in another 7 projects (4 European, two industrial and one TSB funded). In 2011 he was Principal Investigator in the SBRI-funded TRIDS project (Tracking Real Time Intelligence in Data Streams) on building an infrastructure for the analysis of social streams. He is currently principal investigator in the EPSRC-funded project RAnDMS (Real time Analysis of Digital Media Streams) on models and algorithms for data and visual analytics designed to enhance situation awareness through the detection and understanding of events in large scale using heterogeneous and dynamic data sources. He is currently the proposed Director of the European Proposal WeSenseIt (currently in the negotiation phase) on creating Citizen Observatories for the water cycle using social and physical sensors. He has published more than 100 papers. He is part of the editorial board of the International Journal on "Web Semantics", the major International Journal on the Semantic Web. He holds a PhD from the University of East Anglia and a Laurea from the University of Torino, Italy.

Presentation slides (to appear).

April 11th 2012, 13:15 (John von Neumann, IDA)

Presenters: Zlatan Dragisic and Valentina Ivanova

Title: Ontology debugging

Abstract: With the increased use of ontologies and ontology mappings in semantically-enabled applications such as ontology-based search and and data integration, the issue of detecting and repairing defects in ontologies and ontology mappings has become increasingly important. These defects can lead to wrong or incomplete results for the applications. This talk will include three parts. In the first part we discuss the different kinds of defects and exemplify their influence in semantically-enabled applications. In the second part we focus on taxonomies, the currently most used kind of ontologies, and present a unified framework for debugging the is-a structure of and mappings between taxonomies. We also present an implemented system RepOSE, that supports a domain expert to detect and repair missing and wrong is-a relations and mappings. In the third part we discuss the problem of debugging missing is-a relations in more expressive ontologies.

Presentation slides (to appear).

March 14th 2012, 13:15 (John von Neumann, IDA)

Presenter: Jody Foo

Title: Semantic Technologies in Computational Terminology Management

Abstract: Computational Terminology Management (CTM) and Computational Terminology (CT) are areas where computational methods are applied to terminology related tasks. CTM can be divided into three areas: Terminology Creation, Terminology Management, and Terminology Use. There are many semantic technologies and methods which can be shared between CTM and other fields such as Ontology Engineering and Information Retrieval. In this talk I will give a brief overview of the main principles within CTM, followed by an overview of term extraction methods which have been used in Computational Terminology. Finally, I will give some examples of potential areas of research collaboration.

Presentation slides.

February 8th 2012, 13:15 (John von Neumann, IDA)

Presenter: Fredrik Heintz

Title: Integrating Stream Reasoning in Robotic Systems Using Semantic Technologies

Abstract: In this talk I will present our ongoing work on logic-based reasoning over streams of data and the integration of the reasoning in robotic systems through the use of semantic technologies.

For autonomous systems such as unmanned aerial vehicles to successfully perform complex missions, a great deal of embedded reasoning is required at varying levels of abstraction. This reasoning is fundamentally incremental in nature. New observations become incrementally available over time and there is a need to reason about this information with minimal delay as it becomes available. Representing the information as streams captures its incremental nature. The continuous and incremental reasoning over these streams, stream reasoning, allows the system to react to rapid changes in the environment.

When integrating stream reasoning in a robotic system there is a need to connect the symbols used in the reasoning to concrete data streams in the robotic system. Traditionally this connection has been explicit, requiring a specification of the corresponding stream for each symbol. To remove this requirement, we would like to match symbols to streams by reasoning about the meaning of the symbols and the content of the streams. This is achieved by creating an ontology representing the application domain and allowing the robotic system to associate ontological concepts to its streams. Given a set of symbols we can then use the ontology to find those streams which contain the required data.

The system has been implemented and tested in a small UAV application.


From the discussion: Slide with links to Semantic Web applications is here.

January 17th 2012, 13:15 (John von Neumann, IDA)

Presenter: Eva Blomqvist, IDA-HCS

Title: Ontology Engineering for the Semantic Web and Beyond...

Abstract: In the seminar I will first introduce the vision of the semantic web and its current state, as well as my view on ontologies. Then I will proceed to discuss my own research in the field, mainly focused on ontology engineering for the semantic web. At a stage when more and more web developers start creating ontologies, they need tools to understand and pick the right solution for their problem, without having to study the underlying logic in detail. This problem has given rise to the Ontology Design Pattern (ODP) movement (c.f. design patterns in Software Engineering). ODPs come in many different flavours. Initially they were logical patterns, i.e. completely domain independent, however, nowadays there are also domain-specific ODPs and patterns for several other tasks in the ontology life cycle. Ontology engineering methodologies have also been the focus of quite a bit of research during the past decade, and as a reaction against the classical waterfall-style methodologies more recently agile ontology engineering methodologies have been introduced. I will briefly introduce one of them, named eXtreme Design (XD), and some studies made on the methodology. Finally, I will end with a small outlook towards my future research at IDA.


Page responsible: Eva Blomqvist
Last updated: 2018-11-28