ICT, FET Open
LIFT ICT-FP7-255951
Using Local Inference in Massively Distributed Systems
Collaborative Project
D 6.4
Project Workshop
Contractual Date of Delivery:
31.03.2013
Actual Date of Delivery:
08.05.2013
Author(s):
Minos Garofalakis (TUC), Antonis
Deligiannakis (TUC), Michael May (FHG),
Christine Kopp (FHG) , Michael Kamp
(FHG)
Institution:
FHG
Workpackage:
WP 6
Security:
PU
Nature:
R
Total number of pages:
6
Project coordinator name: Michael May
Project coordinator organisation name:
Fraunhofer Institute for Intelligent Analysis
and Information Systems (IAIS)
Schloss Birlinghoven, 53754 Sankt Augustin, Germany
URL: http://www.iais.fraunhofer.de
Revision: 2
Abstract:
This document is the LIFT deliverable of WP6 for the third review period
(01.10.2012 – 30.09.2013). The document contains an overview of two projectrelated workshops to be held in August 2013.
Revision history
Administration Status
Project acronym: LIFT
Document identifier:
Leading Partner:
Report version:
Report preparation date:
Classification:
Nature:
Author(s) and contributors:
Status:
x
ID: ICT-FP7-255951
D 6.4 Project Workshop
(01.10.2012 – 30.09.2013)
FHG
2 (workshop program and details added)
10.12.2013
PU
REPORT
Minos Garofalakis (TUC), Antonis Deligiannakis (TUC),
Michael May (FHG), Christine Kopp (FHG), Michael Kamp
(FHG)
Plan
Draft
Working
Final
Submitted
Copyright
This report is © LIFT Consortium 2013. Its duplication is restricted to the personal use
within the consortium and the European Commission.
www.lift-eu.org
ii
D6.4 – Project Workshop
Minos Garofalakis, Antonis Deligiannakis, Michael May, Christine Kopp
1 Introduction
In year three of LIFT two project-related workshops will take place that are colocated with major scientific conferences in the area of data management and
data mining. The first workshop, Big Dynamic Distributed Data, is co-located with
VLDB 2013 and is organized by TUC. It focuses on the core topic of LIFT, namely
complex computations over partitions of massive streaming data. The second
workshop, Ubiquitous Data Mining, is jointly organized by João Gama and FHG. It
is already the third workshop in its series and has in 2013 a special focus un
distributed streaming data.
2 First International Workshop on Big Dynamic Distributed Data
(BD³ 2013)
Date and Location
August 30th, 2013, Trento, Italy (in conjunction with VLDB 2013)
Chairs
 General Chairs:
 Program Chairs:
Minos Garofalakis, Antonios Deligiannakis
Graham Cormode, Assaf Schuster, Ke Yi
Website
http://www.softnet.tuc.gr/bd3
Program
08:30 - 09:00 Session 1: Invited Talk
Title: Streaming Balanced Partitioning of Massive Scale Graphs
Speaker: Milan Vojnovic, MSR Cambridge
09:00 - 10:00 Session 1: Keynote Speech
Title: Coding Theory for Large-Scale Storage
Speaker: Alex Dimakis, UT Austin
10:00 - 10:30 Coffee break
10:30 - 12:00 Session 2: Distributed Monitoring
Safe-Zones for Monitoring Distributed Streams
Daniel Keren (Haifa), Guy Sagy (Technion), Amir Abboud (Technion), Izchak
Sharfman (Technion), Assaf Schuster (Technion), David Ben-David (Technion)
2
Communication-Efficient Distributed Online Prediction using Dynamic Model
Synchronizations
Mario Boley (Fraunhofer IAIS), Izchak Sharfman (Technion), Daniel Keren
(Haifa), Michael Kamp (Fraunhofer IAIS), Assaf Schuster (Technion)
Communication-efficient Outlier Detection for Scale-out Systems
Moshe Gabel (Technion), Daniel Keren (Haifa), Assaf Schuster (Technion)
12:00 - 13:30 Lunch
13:30 - 15:30 Session 3: CEP and Graphs
Elastic Complex Event Processing under Varying Query Load
Thomas Heinze (SAP AG), Yuanzhen Ji (SAP AG), Yinying Pan (SAP AG), Franz
Josef Grüneberger (SAP AG), Zbigniew Jerzak (SAP AG), Christof Fetzer (TU
Dresden)
Adaptive Selective Replication for Complex Event Processing Systems
Franz Josef Grüneberger (SAP AG), Thomas Heinze (SAP AG), Pascal Felber
(Universite de Neuchatel)
Dynamic Partitioning of Big Hierarchical Graphs
Vasilis Spyropoulos (Athens University of Economics and Business), Yannis
Kotidis (Athens University of Economics and Business)
Scalable and Robust Management of Dynamic Graph Data
Alan Labouseur (SUNY -- Albany), Paul Olsen (SUNY -- Albany), Jeong-Hyon
Hwang (SUNY - Albany)
15:30 - 16:00 Coffee break
16:00 - 18:00 Session 4: Stream Processing
Towards Elastic Stream Processing: Patterns and Infrastructure
Kai-Uwe Sattler (TU Ilmenau), Felix Beier (TU Ilmenau)
Task Graphs of Stream Mining Algorithms
Sayaka Akioka (Meiji University)
Large-scale Online Mobility Monitoring with Exponential Histograms
Christine Kopp (Fraunhofer IAIS), Michael Mock (Fraunhofer IAIS), Odysseas
Papapetrou (Technical Univ. of Crete), Michael May (Fraunhofer IAIS)
Multi-Stage Malicious Click Detection on Large Scale Web Advertising Data
Leyi Song (East China Normal University), Xueqing Gong (East China Normal
University), Xiaofeng He (East China Normal University), Rong Zhang (East China
Normal University), Aoying Zhou (East China Normal University)
18:00 - 18:10 Closing Remarks
3
Workshop Proceedings
http://ceur-ws.org/Vol-1018/
Workshop Description
As the amount of streaming data produced by large-scale systems such as
environmental monitoring, scientific experiments and communication networks
grows rapidly, new approaches are needed to effectively process and analyze
such data. There are several promising directions in the area of large-scale
distributed computation, that is, where multiple computing entities work together
over partitions of the massive, streaming data to perform complex computations.
Two important paradigms in this realm are continuous distributed monitoring
(i.e., continually maintaining an accurate estimate of a complex query), and
distributed and cluster-based systems that allow the processing of big, streaming
data (e.g., IBM System S, Apache S4, and Twitter Storm).
The aim of the BD3 workshop is to bring together computer scientists with
interests in this field to present recent innovations, find topics of common
interest and to stimulate further development of new approaches to deal with
massive dynamic and distributed data.
Topics of interest include (but are not limited to):












Novel architectures for BD3
Extensions to existing models for BD3
Algorithms for mining and analytics for BD3
Query processing in BD3
Efficient communication protocols for BD3
Languages and structures for BD3
Theoretical basis and hardness for BD3
Engineering case-studies in BD3
Position papers on challenges and new directions in BD3
Privacy issues in BD3
Energy efficiency and reliability in BD3
Scheduling and provisioning issues in BD3
The 1st International Workshop on Big, Dynamic, Distributed Data (BD3) took
place on Friday 30/8, in conjunction with VLDB'2013 in Riva Del Garda, Italy.
(VLDB is the top international research forum on data management systems.)
The workshop received 15 submissions out of which 11 papers were selected for
presentation. Four of the accepted papers were, in fact, by LIFT authors, so LIFT
work took center stage at the workshop. Professor Alex Dimakis from UT-Austin
delivered a very interesting keynote talk on novel applications of coding theory in
robust distributed storage. VLDB handled workshop registrations centrally, so
people registered for all workshops on that day - there were 8 concurrent
workshops taking place on 30/8. The BD3 workshop was very well attended, with
30-40 participants in the workshop throughout the day, and several interesting
technical discussions taking place both during the sessions and the coffee breaks.
4
3 Workshop on Ubiquitous Data Mining (UDM 2013)
Date and Location
August 3rd-5th, 2013, Beijing, China (in conjunction with IJCAI 2013)
Chairs
 João Gama, Michael May, Nuno Marques, Paulo Cortez
Website
http://www.liaad.up.pt/udm
Program
09:00 -10:00 Invited Talk
Title: Exploiting Label Relationship in Multi-Label Learning
Speaker: Zhi-Hua Zhou
10:00 – 12:25 Paper Session 1
Predicting Globally and Locally: A Comparison of Methods for Vehicle Trajectory
William Groves, Ernesto Nunes and Maria Gini
On Recommending Urban Hotspots to Find Our Next Passenger
Luis Moreira-Matias, Ricardo Fernandes, Joao Gama, Michel Ferreira, João
Mendes-Moreira and Luis Damas
Road-quality classification and bump detection
smartphones
Marius Hoffmann, Michael Mock and Michael May
with
bycicle-mounted
Cicerone: Design of a Real-Time Area Knowledge-Enhanced Venue Recommender
Daniel Villatoro, Jordi Aranda, Marc Planagumà, Rafael Gimenez and Marc
Torrent-Moreno
Visual Scenes Clustering Using Variational Incremental Learning of Infinite
Generalized Dirichlet Mixture
Wentao Fan and Nizar Bouguila
12:30 – 13:30 Lunch
13:30 – 14:30 Invited Talk
Title: NIM: Scalable Distributed Stream Processing System on Mobile Network
Data
Speaker: Wei Fan
14:30 – 17:00 Paper Session 2
5
Learning Model Rules from High-Speed Data Streams
Ezilda Almeida, Carlos A. Ferreira and João Gama
Simultaneous segmentation and recognition of gestures for human-machine
interaction
Harold Vasquez, Luis Enrique Sucar and Hugo Jair Escalante
Ubiquitous Self-Organizing Maps
Bruno Silva and Nuno C. Marques
Simulating Price Interactions by Mining Multivariate Financial Time Series
Bruno Silva, Luis Cavique and Nuno C. Marques
Trend template: mining trends with a semi-formal trend model
Olga Streibel
17:00 Closing
Workshop Proceedings
http://www.liaad.up.pt/udm/workshop-proceedings/at_download/file
Workshop Description
Two major technological
environment:


evolutions
modify
our
relationship
with
our
Widely available and cheap computer power. Simple objects that surround
us are gaining sensors, computational power, and actuators, and are
changing from static, into adaptive and reactive systems.
The explosion of networks of all kinds offers new possibilities for the
development and self-organization of communities.
The new characteristics of data reflect a World in Movement:






Time and space. The objects of analysis exist in time and space. Often
they are able to move.
Dynamic environment. These objects exist in a dynamic and unstable
environment, evolving incrementally over time.
Information processing capability. The objects are endowed with
information processing capabilities.
Locality. The objects never see the global picture - they know only their
local spatio-temporal environment.
Real-Time. The models have to evolve incrementally in correspondence
with the evolving environment.
Distributed. The object will be able to exchange information with other
objects, thus forming a truly distributed environment.
Ubiquitous Data Mining (UDM) uses Data Mining techniques to extract useful
knowledge from data with these characteristics. The goal of this workshop is to
convene researchers (from both academia and industry) who deal with
6
techniques such as: decision rules, decision trees, association rules, clustering,
filtering, learning classifier systems neural networks, support vector machines,
preprocessing, post processing, feature selection, visualization techniques, etc.
for UDM and related themes.
Topics include but are not restricted to:














Adaptive Data Mining
Distributed Data Mining
Distributed Data Streams
Grid Data Mining
Learning in Ubiquitous environments
Learning from Sensor Networks
Learning from Social Networks
Visualization Techniques for UDM
Incremental On-line Learning Algorithms
Single-Pass and Scalable Algorithms
Learning in distributed neural network systems;
Real-Time and Real-World Applications
Resource-aware UDM
Theoretical frameworks for UDM
The Workshop on Ubiquitous Data Mining took place on Saturday 3/8, in
conjunction with IJCAI’2013 in Beijing, China. The workshop received 16
submissions out of which 8 papers were selected for presentation by a scientific
committee. The UDM workshop was well attended, with around 25 participants in
the workshop throughout the day, and several interesting technical discussions
taking place both during the sessions and breaks.
7