Proud contributor to "Being Profiled: Cogito Ergo Sum":

A book in the occasion of 10 years of ‘PROFILING THE EUROPEAN CITIZEN’, an EU funded project that covered the downsides of profiling very early on, initiated by Prof. Mireille Hildebrand and colleagues.

Want to attend the presentation of the first copy to Paul Nemitz, Principal Advisor in DG JUSTICE at the European Commission, and
a panel discussion between representatives from Microsoft, UCL and SAS?

brusselsprivacyhub.eu/events/1

Announcement of exposition of paintings by the Utrecht based artist John Hoy, in Hooghiemstra.

@djoerd want to travel via HH / spinque on the way to EUT in Merleyn?

Data on tracking collected by whotracks.me (Cliqz, owns Ghostery):

whotracks.me/

Google, comscore, Facebook, Amazon, Criteo, Yandex and Cloudflare in the top stats.

Provenance: whotracks.me/blog/where_is_the

The little joys of life: three "rel=me" backreferences working correctly to link my identity on three ActivityPub services in the Fediverse!

ICT with Industry 2019, 60 researchers working on 5 case studies.

Held at the Lorentz Center, 21-25 January 2019.

Organized by Lynda Hardman (CWI) and Suzan Verberne (Faculty of Science - Universiteit Leiden) with NWO and 5 companies:

- FD Mediagroep
- ICTgroup
- KB (National Library of the Netherlands)
- Persgroep
- RTL Nederland

Bye see you in Cologne for ECIR, Glasgow for CHIIR, Paris for SIGIR or DIR 2019 in Amsterdam!

Tnx @suzan!

Livestream of MdR's Tony Kent Strix Award acceptance speech, from London to Leiden.

Seyyed Hadi Hashemi presents two papers on evaluation and analysis of personal assistents (Alexa, Google Home etc.)

Discusses segmentation, task success prediction, and intent-sensitive embeddings for query representation.

Dataset with 500k sessions with 181 intents (Microsoft data, closed...)

The former ILPS gang taking over medialand in NL, ready at the balcony for the livestream 🍿

Research by @djoerd his MSc students. A project presented by Jan Trienes, on "who to follow" on Mastodon.

Results of scientometric analysis of 21.2M WoS papers published 2000-2017

Identified the classic, the CHIIR, and the bio-informatics perspectives on IR. (Probably some bias toward journal papers due to WoS.)

Uses VOSviewer and the Leiden Algorithm (Traag et al., 2018)

Ludo discusses the similarity between Pagerank vs. the classic Pinski and Narin scientometrics paper in IP&M (1976).

BTW:
For history of Pagerank before Pagerank, see also arxiv.org/abs/1002.2858

Second keynote by CWTSs Ludo Waltman on the intersection between Scientometrics and IR

Especially:
Scientific Document Retrieval.

Identifying the relevant set of articles (for the domain of interest) is the starting point of any scientometric analysis, so shared interest!

Users often decision takers interested in an overview of domain instead of specific articles.

Co-citation analysis White and McCain 1998 illustrates historically strong connection even if separate subdomains

Challenges in Legal IR mentioned by legal professionals in Ginneke's study.

Ginneke Wiggers presents recent work in professional search carried out with Legal Intelligence.

Focused questionnaire to collect pairwise evidence of preferences related to relevance factors like recency, legal hierarchy, source authority, etc. (building on Opijnen and Santos 2017).

Document type and level of depth/detail are very important factors in the overall relevance assessment.

Recent work preview: comparing info specialist to practitioners, seem to have different preferences.

program and proceedings:
dir2018.nl/programme

We just had Felipe Moraes from Delft present a nice user study combining search and learning: search as a learning activity. Details in the proceedings!

SearchX collaborative search system:
FelipeMoraes.github.io/searchx
or fmoraes.nl/searchx/

Show more
Mastodon

The "unofficial" Information Retrieval Mastodon Instance.

Goal: Make idf.social a viable and valuable social space for anyone working in Information Retrieval and related scientific research.

Everyone welcome but expect some level of geekiness on the instance and federated timelines.