Ref. Ares(2016)5296138 - 14/09/2016
Commission of the European Communities
Date: 15.06.2016
Joint Research Centre - Ispra Establishment
Mission No: DF-16-1221405
Mission Report
Name
Division/Action/Project
GlobeSec/
EMM
Object of the mission
The objective of the mission was to deliver an invited talk at the
Deggendorf Forum for Digital
Information Analysis (DFDDA; http://www.dfdda.de) and to attend the yearly conference.
Place, date, organizations visited and contacted persons
(Please continue on a separate sheet, if necessary)
Deggendorf Forum zur Digitalen Datenanalyse
Garching, Germany, 9-10 June 2016
Institute for Advanced Study, Technical University Munich.
Other Participants
Distribution List
Of the establishment
Other services of the commission
Obligatory(*) +
.
(*), JRC
(*), JRC
(*), JRC
xxxxxxxxxxxxxxxxxxxx@xxx.xx.xxxxxx.xx (*), JRC
GlobeSec Unit, JRC xxxxxxxxxxxxx@xxx.xx
Summary
The objective of the mission was to deliver an invited talk at the
Deggendorf Forum for Digital
Information Analysis (DFDDA; http://www.dfdda.de) and to attend the yearly conference, which took
place in Garching, near Munich, Germany. DFDDA mostly targets auditors and it includes new
technologies that may become important for their work. For the program of the event, see
http://www.dfdda.de/dfdda2016/dfdda20160.html.
The one-hour JRC presentation was one of seven presentations at the forum. It had the title ’Supporting
the analyst by turning multilingual text into structured data’. In the presentation, I gave an overview of
EMM and OSINT Tools, highlighting the benefits of large-scale data analysis, showing various ways
to visualise data, and making users aware of the dangers of fully-automatic analysis. I aimed to address
the specific needs of auditors and I tried to show what an EMM-like analysis can do for this type of
users and what it cannot do. Finally, I presented the multilingual linguistic resources created as by-
products of EMM and downloadable from the JRC’s Science Hub pages, as well as from the EU Open
Data Portal, with a special focus on
JRC-Names and the
JRC Eurovoc Indexer JEX.
The talk was received extremely well. There were many questions regarding functionality, availability
of the software and the possible usage of the linguistic resources.
The DFDDA event was useful to get a good insight into the needs of auditors. The presentation I found
most useful for our own purposes was the talk by
from Passau
University (Germany), who talked about combining formal and distributional semantics to represent
words and texts, showing the benefit for information retrieval and question answering. While in formal
semantics, meaning aspects of concepts are intellectually identified and described (e.g. a ‘dog’ has a
name, a colour, a race, a size, etc.), the individual values for each of these meaning aspects can be
found in large amounts of texts, using distributional methods. The commonalities and the differences
between the concepts ‘apple’ and ‘pear’ can thus be identified, for instance, by searching which
adjectives and which verbs (subject of, object of) are found with the equivalent words.
also
presented word clouds around the country name ‘Iraq’ (associated words) and how they changed over
time. This latter kind of observation could also be useful as an analysis aspect to be displayed in EMM.
, 15 June 2016