Drift-a-LOD2017: Detection, Representation and Management of Concept Drift in Linked Open Data

Time: 
Monday, September 11, 2017 - 09:00 to 17:30
Place: 
The Meervaart (Room 5)

Chairs: Sándor Darányi (University of Borås, Sweden), Laura Hollink (Centrum Wiskunde & Informatica, The Netherlands), Albert Meroño Peñuela (Vrije Universiteit Amsterdam, The Netherlands), Efstratios Kontopoulos (Center for Research & Technology Hellas, Greece)

PROGRAM

09.00-09.05 : Workshop Welcome
09.05-10.00 : Keynote 1 by Antske Fokkens: "Drifting distributions? Possibilities and risks of using distributional semantics for studying concept drift"
10.00-10.30 : "Exploring Concept Representations for Concept Drift Detection" by Oliver Becher, Laura Hollink and Desmond Elliott.
 
10.30-11.00 : Coffee break
 
11.00-12.00 : Keynote 2 by Javier Fernández: "LOD is all about evolution: Querying and Managing evolving Linked Open Data"
12.00-12.30 : "A Study of Intensional Concept Drift in Trending DBpedia Concepts" by Albert Meroño-Peñuela, Efstratios Kontopoulos, Sándor Darányi and Yiannis Kompatsiaris
 
12.30-13.30: Lunch
 
13.30-14.00: Industry presentation: "Modeling, Measuring and Exploiting Concept Drift in the Labour Market Domain" by Panos Alexopoulos and Spyretta Leivaditi
14:00-14:30: "Researching the Presence of the Past in Diachronic Corpora" by Alex Olieman, Kaspar Beelen and Jaap Kamps
14.30-15.00: Discussion
 

KEYNOTES

 
Drifting distributions? Possibilities and risks of using distributional semantics for studying concept drift.
Recent work in NLP and digital humanities has seen a variety of study where distributional semantics is used to study changes in meaning, but can these technologies be used this way? Whether and how distributional semantic technologies can be used to study concept drift depends on three things: first, how concept drift is defined exactly, second, whether this concept can be translated to expressions that provide insight into potential change and third, the availability of sufficient data for creating reliable distributional semantic models. I will show that reliability and stability form a significant challenge when working with diachronic datasets. I will dive deeper into the challenges of using distributional semantics for studying data drift by examining a use case on changes of the concept of racism using results from Pia Sommerauer's master thesis.
 
 
Javier Fernández
 
LOD is all about evolution: Querying and Managing evolving Linked Open Data
The steady adoption of Linked Data in recent years has led to a significant increase in the number and volume of RDF datasets. However, in the absence of a central control mechanism, this huge knowledge base is ephemeral: datasets constantly appear, change and disappear. In the first part of the talk, after introducing the general challenges emerging in a Big Semantic Data scenario, we will inspect the challenges of representing and querying evolving semantic data. Then, we will present different modeling strategies, RDF indexes and practical tools to cope with RDF versions, allowing cross-time queries to understand and analyse the history and evolution of dynamic datasets. 

 

ABOUT THE WORKSHOP

The continuous growth of the Linked Open Data (LOD) cloud is extending to various new domains. In many of these, facts change continuously: political landscapes evolve, medical discoveries lead to new cures, artists form new collaborations. In terms of knowledge representation, we observe that instances change their roles, new relations appear, old ones become invalid, and classes change both their definition and member-instances.

The evolution of LOD poses concrete new challenges to stakeholders: data publishers need to detect changes in the real world and capture them in their datasets; users and applications need automated tools to adapt querying over diachronic datasets; knowledge engineers want to understand modelling practices behind ontology changes; linguists study drift in the meaning of words. As a continuation of last year’s successful Drift-a-LOD, this workshop seeks to form a community of researchers and practitioners working on detecting, representing and managing concept drift in and for LOD.

Drift-a-LOD’17 will bring together different communities that define, identify and manage the dynamics of concepts in their knowledge bases using various domain-specific methods (statistical inference, symbolic reasoning, natural language processing, etc.), leveraging the LOD cloud as a data source or as a result publishing platform.
 

TOPICS

Topics of interest include, but are not restricted to:

  • detecting and predicting concept drift (using any method, incl. reasoning, data mining, word embeddings, NLP)
  • representation of concept drift
  • reasoning, querying, machine learning in the presence of evolving knowledge and drifting concepts
  • theoretical explanations of drift dynamics
  • visualization and presentation of evolving knowledge
  • ontology evolution and concept drift
  • empirical studies of how concepts drift
  • evaluation of concept drift detection methods
  • applications working in the presence of concept drift
  • frameworks addressing concept identity over time
     

IMPORTANT DATES

July 4, 2017: deadline to submit papers
July 11, 2017: extended deadline to submit papers
August 7, 2017: notifications to authors
August 14, 2017: camera ready versions
September 11, 2017: workshop

SUBMISSION GUIDELINES

We invite full papers and short papers. Contributions should follow the ACM ICPS guidelines for formatting and must not exceed 8 pages in length for full papers and 4 pages for short papers, including references and optional appendices. The layout templates can be found here: http://www.acm.org/sigs/publications/proceedings-templates. Papers should be submitted through the EasyChair submission system at https://easychair.org/conferences/?conf=driftalod2017

Contributions may be accepted as either long of short presentations depending on quality, novelty, and potential to stimulate a discussion at the workshop. Accepted contributions will be published on the CEUR-WS website (or equivalent).

At least one author needs to register for the workshop. Day passes for the workshop are available at 40 euro (excl. VAT).
 

ORGANISING COMMITTEE

PROGRAM COMMITTEE

  • Astrid van Aggelen, Centrum Wiskunde & Informatica, NL
  • Antonis Bikakis, University College London, UK
  • Irini Fundulaki, Foundation for Research and Technology - Hellas (FORTH), GR
  • Tomi Kauppinen, Aalto University, FI
  • Nikolaos Lagos, Xerox Research Centre Europe
  • George Meditskos, Center for Research & Technology Hellas (CERTH), GR
  • Carlo Meghini, Consiglio Nazionale delle Ricerche (CNR), IT
  • Francesco Osborne, Knowledge Media Institute, UK
  • Thanos Stavropoulos, Center for Research & Technology Hellas (CERTH), GR
  • Jannik Stroetgen, Max-Planck-Institut für Informatik, DE
  • Ilaria Tiddi, Knowledge Media Institute, UK
  • ...
  • To be extended... (let us know if you would like to be a member of the PC!)