Welcome to our third Newsletter

I N S I D E T H I S

I S S U E :

Welcome 1

1st MOVING

plaorm proto-

type

User require-

ments and user

studies

History view 2

Learning-how-

to-search widg-

URank visualisa-

on

Concept Graph 3

New MOVING

logo

MulEdTech 4

MOVING

presentaons in

events

MOVING paper

awarded @ EICS

2017!

MOVING at OER

Congress

MOVING in 2

seminars

Brief news: re-

cent and up-

coming events

Contact details 6

1st MOVING platform prototype is publicly available!

D E C E M B E R 2 0 1 7

I S S U E 3

The rst prototype of the MOVING plaorm is now available to the public! The MOVING plaorm

enables its users to improve their informaon literacy by training how to exploit data and text mining

methods in their daily research tasks. The MOVING search engine provides scalable real-me search,

supports mulple document types, dierent le formats and dierent programming languages. Faceted

search allows to retrieve various kinds of documents such as scienc arcles, books, video lectures,

and metadata. Graph visualisaon highlights relaons among documents and related enes (authors,

organisaons, etc.) and oers an alternave way of exploring search results. Nevertheless, a classical

search list is sll also featured. For example, the user can click on an author name and retrieve all docu-

ments authored by this person. To ensure a smooth user experience, dedicated tools allow us to sepa-

rate dierent authors with the same name or connect dierent versions of the same document. The

Adapve Training Support provides illustrated feedback to the user in order to help her get familiar

with the plaorm and all its features (for example by recommending features the user has not tried so

far). You can try out the plaorm at hps://moving.mz.test.tu-dresden.de/. Don’t forget to send us

any feedback that you might have!

We have already entered the second half of the second year of our project. We are pleased to an-

nounce that the rst prototype of the MOVING plaorm, based on the rst set of user requirements,

has become publicly available! In parallel to this, we connued the development of individual soware

components of the plaorm such as video analysis, scienc gures search and hypothesis tesng for

web interacon paerns, and a number of scienc publicaons detailing the results were further pro-

duced. We also iniated the implementaon of the two user studies, regarding the research on busi-

ness informaon by public administrators, and managing and mining research informaon by research-

ers, and we will soon start their evaluaon. Finally, the MOVING consorum connued in this period to

parcipate in several scienc and industrial events.

The MOVING search page

P A G E 2

The history

view of

MOVING

plaorm

Tools - Demos - Results

User requirements and initial implementation of user

studies

The user requirements are the result of the analyses performed during the rst year of the project. Most

of these requirements have been implemented already in the MOVING plaorm. During our last project

meeng, the requirements list was extensively discussed and re-priorized, so that most important re-

quirements are made available in the rst public release of the plaorm. The search and faceted search

funconalies have been almost completely implemented, and so are the data sources that the queries

rely on to provide results. The concept graph, a very important user requirement, is implemented and

ready to use. In order to test the user acceptance and evaluate the use of the plaorm, we will conduct

user studies, taking place during the last months of the project’s second year. The evaluaons of these

studies will provide insight on what needs to be improved regarding the implementaon of the require-

ments. That will then lead to a beer user experience, which we will test and evaluate again to check for

improvements. These studies are key to the funconalies of the plaorm and future implementaons.

Learning-how-

to-search

widget of the

MOVING

plaorm

A rst version of the “learning-how-to-search” widget is im-

plemented in the MOVING plaorm. The MOVING user can

use the widget and become a search expert. The widget

provides feedback about the user’s search behavior in order

to nudge the user to experiment with dierent search func-

onalies and corresponding visualisaons available on the

plaorm. For this, it presents a bar chart showing usage sta-

scs for each feature, allowing the user to get a quick over-

view of the own behavior. Occasionally, the widget prompts

quesons targeng dierent search-related topics, like

poinng to less used or new features and at a later stage

asking about the most sasfactory or favorite features. The

feedback textbox gives the user the opportunity to write

down his/her thoughts immediately, for revising them at a

later point in me.

“Learning-how-to-search” widget

The history view in the search results of

the MOVING plaorm shows the dierent

versions of laws and regulaons, as show

in the gure on the right. The history view

can help compliance ocers to track the

evoluon of these documents over me

and refer to a specic version, if needed.

It disnguishes dierent kinds of events,

such as paral alteraon and rewrite. The

laws and regulaons dataset was provided

by Wolters Kluwer within the H2020 pro-

ject ALIGNED.

The history view of the search results in the platform

The history view for the word “Gesetz” (“Law”)

User

requirements

and user

studies status

Concept graph visualisation

P A G E 3 I S S U E 3

The Concept graph is a node-link representaon that oers a

novel way to nd and explore data in the MOVING plaorm.

It shows the connecons between the search results and

related features (enes), such as authors, their aliaons,

and publicaon dates. Addional enes, such as subjects,

keywords or locaons may be added in the future. The user

can explore the result set by navigang along relaonships

connecng documents and enes. Aer entering a query in

the search eld, the 12 most relevant results will be dis-

played as nodes in a circular fashion. Hovering over a node

will display informaon about the document and double-

clicking will open the document in a new browser tab. Click-

ing on a node will expand the graph by displaying directly

connected nodes, such as the authors of the document or

the publicaon year. By expanding a node, both new con-

nected nodes and their connecons to other already visible

nodes are added to the graph. The user can quickly navigate

through a network of relevant nodes starng from the re-

sults of the entered search query. For example, by expanding

a document node the user could discover a connecon to

other documents, which were published by the same author

or year (or both).

URank visualisation

uRank is an interacve web-based tool implemented in the

MOVING plaorm, that supports users during exploratory

search. The approach combines lightweight text analycs

and a stacked-bar-based visualisaon to convey a content-

based ranking of documents in a search result set. The key of

this approach lies in its dynamic nature, supported by a user-

driven method for updang the ranking visualisaon that

allows users to rene their search interest as they evolve

during exploraon. The documents and their keywords are

passed to the uRank user interface, which consists of three

main components: i) The Tag Box, which oers a summary

of the document collecon by represenng extracted key-

words as tags. ii) The Query Box, which is a container where

the user drops keywords of parcular interest. iii) The Docu-

ment List, which shows the hits sorted according to user-

selected keywords. Each hit includes the document tle, its

ranking posion informaon, and a stacked bar chart de-

picng relevance scores in terms of user-selected keywords

(i.e. keywords in the Query Box). Changes in the ranking

representaon originate from manipulang keyword tags

within the Query Box in three dierent ways: i) addion by

dragging and dropping tags from the Tag Box, ii) weight

tuning through keyword-sliders, and iii) deleon. As the

user selects terms of interest, the ranking visualisaon

brings related documents to the top and pushes down less

relevant ones. Finally, the uRank interface includes a topi-

cal tag cloud that summarises the content of the result set,

and provides means for dynamic, user interest-driven re-

ranking and exploraon of search results.

The concept graph with the expanded nodes

The URank tool in the MOVING plaorm

The expanded

ring-menu of

a node and

the created

subgraph

MOVING-

organised

workshop at

ACMMM 2017

P A G E 4

The MOVING-organised workshop MulEdTech2017 at ACMMM 2017 was successfully concluded on

October 27th in Mountain View, California, USA. A keynote talk and four oral paper presentaons were

delivered at the workshop and sparked very interesng discussions on mulmedia-based educaonal

and knowledge technologies, and parcularly on visualisaon technologies on mulmedia content avail-

able in specialised learning plaorms, the Web, mobile devices and/or social networks for supporng

personalised and adapve e-learning and training. You can nd the complete set of presentaons in the

workshop’s website. The workshop’s proceedings have also been made available by the ACM, and can

be accessed at hps://dl.acm.org/citaon.cfm?id=3132390 .

Successful organisation of MultiEdTech 2017

Communication and dissemination activities

Evoluon, rather than revoluon, was our guiding principle for de-

signing the new logo of the project. The new logo, pictured on the

right, looks modern and fresh, and at the same me is unmistakably “MOVING”: with its colors and de-

sign, it maintains the visual idenfy that the project has been building since day one.

New MOVING

logo

as a Cloud-Based Decision Sup-

port Service” at MOPGP’17. The

presentaon contained the ex-

amples from the Delphi survey

performed within the MOVING

project, and over 100 experts

from all over the world aended

this session where a long and

vivid discussion took place.

The paper “Mulcriteria decision

planning with ancipatory net-

works”, was successfully present-

ed by Dr. A.M.J. Skulimowski

from PBF at MCDM 2017 confer-

ence. Over 200 top experts from

32 countries took part in the con-

ference, and the paper will be

published as a book chapter in

the MCDM2017 post proceedings

by the CRC Press. Also the paper

“Applicaons of the reference set

method to the priorizaon of

technological strategies of a

knowledge repository” was pre-

sented at SOCA’17. The paper

referred to the elaboraon of

technological hints arising from

the Delphi survey performed in

MOVING. Finally, Prof. Dr. A.M.J.

Skulimowski was invited to give

the talk on “Expert Delphi Survey

MOVING presentations on future prospects and exploit-

ation strategy at MCDM’17, SOCA’17 and MOPGP’17

Query, Aitor Apaolaza and Mar-

kel Vigo were awarded the best

paper award at the conference!

WevQuery is built in the context

of the MOVING project, where a

dashboard for user interacon

analysis provides behavioural

data to other tasks, including the

learning-how-to-search widget

that generates personalised

learning opportunies.

MOVING paper awarded @ EICS 2017!

The paper

“WevQuery:

Tesng Hy-

potheses

about Web

Interacon

Paerns” was

presented by

Aitor Apaolaza at the 9th ACM

SIGCHI Symposium on Engineer-

ing Interacve Compung Sys-

tems (EICS’2017). WevQuery sup-

ports the tesng of hypothesis

about user interacon, by allow-

ing designers to create queries

based on event sequences easily.

This way WevQuery makes com-

plex and low-level web interac-

on events, such as mouse move-

ment and key presses, accessible

to users with lile or none query-

ing experse. A demo was also

presented, allowing the confer-

ence aendees to use Wev-

Query’s interface to create their

queries. For their work on Wev-

New MOVING logo

MOVING paper

awarded @ EICS

2017

Two papers and

a talk at

MCDM’17,

SOCA’17 and

MOPGP’17

Brief news: recent and upcoming events

 A paper on “Content Recommendaon through Semanc Annotaon of User Reviews and Linked Data“ by I. Vagliano, D.

Mon, A. Scherp, M. Morisio was presented at the 9th Internaonal Conference on Knowledge Capture (K-CAP2017), Aus-

n, Texas, United States, 4th-6th December 2017.

 Two talks on “Reranking-based Recommender System with Deep Learning“ by A. Saleh, F. Mai, C. Nishioka, A. Scherp and

on “Word Embeddings for Praccal Informaon Retrieval” by L. Galke, A. Saleh, A. Scherp were given in the workshop on

“Deep Learning in heterogenen Datenbeständen” (at 47. Jahrestagung der Gesellscha fuer Informak) - ws34-dlhd

( 2017 ), Chemnitz, Germany, 25th -29th September 2017.

 The paper “SemRevRec: A Recommender System based on User Reviews and Linked Data”, by I. Vagliano, D. Mon, M.

Morisio was presented in the 11th ACM Conference on Recommender Systems. Como, Italy, 27th-31st August 2017.

 A paper on “Cognive Content Recommendaon in Digital Knowledge Repositories – a Survey of Recent Trends” was pre-

sented in the 16th Internaonal Conference on Arcial Intelligence and So Compung (ICAISC'2017), Zakopane, Poland,

June 11-15, 2017.

 A paper on "Linear Maximum Margin Classier for Learning from Uncertain Data", by C. Tzelepis, V. Mezaris, and I. Patras,

was accepted for publicaon in the IEEE Transacons on Paern Analysis and Machine Intelligence.

More news about these and other MOVING acvies:

hp://moving-project.eu/index.php/category/news/

P A G E 5

MOVING successfully parcipated in the 2nd World OER Congress held in Ljubljana, Slove-

nia, from September 18-20 2017, by organising the satellite event tled “Technologies for

OER and Open educaon: European research projects showcase session”. The MOVING

satellite event was an opportunity for presenng the latest developments in various MOV-

ING-related technologies (video understanding, advanced visualisaon, search technolo-

gies, smart learning etc.) and had about 25 parcipants from dierent sectors (EC, compa-

nies, policy makers), including teachers, researchers, linguists, IPR lawyers, public adminis-

trators etc. You can watch the presentaons online at: hp://videolectures. net/

oercongress2017_satellite _events/. Further to the satellite event, MOVING was also pre-

sent by having a stand with demos and videos at the main congress exhibion hall,

throughout the duraon of the congress. This allowed us to disseminate MOVING to more

than 500 aendees from 111 countries. Most of them were high-level policy makers, repre-

senng the ministry of educaon or other similar enes of their country.

The MOVING project

and its exploitaon

potenal have been

presented by Prof. Dr.

Andrzej M.J. Sku-

limowski (PBF) at the

Expert Seminar on

“Exploitaon of the

MOVING results by academic communies” at the PBF’s

premises in Kraków on June 30, 2017. 12 experts from 7

Polish instuons took part in this event. Also, a seminar for

students and young researchers, including a presentaon of

the MOVING knowledge repository capabilies and a discus-

sion on knowledge repository in the context of decision sup-

port, e-science, business models and forecasng was orga-

nized by the DSS Laboratory of the AGH University of Science

and Technology, the Student Scienc Society on Financial

Modelling, and the PBF at the AGH premises in Kraków on

June 22, 2017. 22 parcipants, mostly PhD students from the

Faculty of Electrical Engineering, Automac Control, Comput-

er Science and Biomedical Engineering, and MSc students

from the Faculty of Mechanical Engineering and Robocs of

the AGH UST, took part in this event that comprised alto-

gether 12 presentaons and discussions.

The MOVING stand

Successful presentations of MOVING and exhibition stand at the

2017 OER congress

MOVING presented in 2 seminars in Kraków

P A G E 6 I S S U E 3

This project has received funding from the European

Union’s Horizon 2020 research and innovaon pro-

gramme under grant agreement No 693092

Project Coordinator: Dr. Vasileios Mezaris

Address: Informaon Technologies Instute (ITI),

Centre for Research and Technology Hellas

(CERTH) 6th Km Charilaou - Thermi Road, 57001

Thermi - Thessaloniki, Greece

Tel: +302311 257770, Fax: +302310 474128

E-mail: bmezaris@i.gr

Full Title: “TraininG towards a society of data-

saVvy informaon prOfessionals to enable open

leadership INnovaon”

Project Idener: H2020 - 693092

Start Date: 1st April 2016

End Date: 31st March 2019

Duraon: 36 months

Contact Details

Project Details

Centre for Research & Technology Hellas

Informaon Technologies Instute

hp://www.i.gr

Ernst & Young GmbH

hp://www.ey.com/Home

Technische Universität Dresden

hps://tu-dresden.de

Know-Center

hp://www.know-center.tugraz.at/

Instut Jožef Stefan

hps://www.ijs.si/ijsw/JSI

ZBW-Leibniz Informaon Centre for Economics

hp://www.zbw.eu/en/

The University of Manchester

hp://www.manchester.ac.uk/

GESIS-Leibniz Instute for the Social Sciences

hp://www.gesis.org/en/instute/

Fundacja Progress and Business

hp://www.pbf.pl

hp://moving-project.eu/

@MOVING_EU