ACM SIGMM Interest List


Options: Use Forum View

Use Monospaced Font
Show HTML Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
"Kraaij, W. (Wessel)" <[log in to unmask]>
Reply To:
Kraaij, W. (Wessel)
Fri, 12 Apr 2013 15:31:02 +0000
text/plain (4 kB) , text/html (12 kB)
Please find below a description of a project for which we seek a PhD candidate.
PhD project "Real time event search"
In many (military) operations the amount of sensor data is too large for effective real-time human evaluation. This project will work towards a system which supports autonomous decentralized processing aiding an operator to overcome his current operator overload.
 Goal of the "Real time event search" project is to effectively execute search strategies on many sensors, where search queries consider multiple objects/agents  including their spatial and/or temporal relationship. Research will be related to the following research questions:
-          How to define of the quality dimensions of the resulting situational awareness system and its components, and what are proper associated metrics?
-          How do user queries about relationships between objects translate to requirements for data analysis, data interpretation and information retrieval modules?
-          What are effective search strategies for information needs addressing objects and their relationships given the uncertainties of available  data analysis, data interpretation and information retrieval components?
-          How can external data resources such as annotated image data and world knowledge encoded in ontologies be utilized to enhance search effectiveness.

This project will be part of the GOOSE program which aims at the development of a system where data analysis, data interpretation and information retrieval components are configured and executed to obtain a many domain, many query, many user and many sensor system. The GOOSE concept has the ambition to provide the capability to search semantically for any relevant information within "all" (including imaging/video) sensor streams, in near real time,  in the entire internet of sensors. The ambition is to realize a search capability for sensor data, which is just as effective and intuitive as using search engines for finding information on the web. Research will utilize the TRECVID MED/MER data sets and evaluation framework.

Promotor of this project will be Prof. Dr. Ir. Wessel Kraaij, affiliated to both Radboud University Nijmegen (ICIS, section intelligent systems)  and TNO. Daily supervision will be provided by Dr. Klamer Schutte, program leader of the TNO GOOSE program. Majority of work will be carried out at TNO The Hague in close collaboration with  TNO Delft and Radboud University Nijmegen.
The "Real time event search" PhD project is fully funded for four years. The successful applicant has
*         a MsC in Computer Science, Artificial Intelligence or equivalent,
*         proven experience with machine learning, data analysis and information retrieval models, both at a theoretical and practical level.
*         good programming skills (e.g. Matlab, Java, Python)
*         excellent written and spoken English language skills (knowledge of Dutch is preferential)
*         excellent communication skills, and the ability to work independently and in a multidisciplinary team.
Given TNO's security policy we only can consider candidates with EU or NATO country nationality.

We accept applications until May 15, or until the position is filled. Please sent your application to Klamer Schutte and Wessel Kraaij. Further information can be obtained by contacting Klamer Schutte, [log in to unmask]<mailto:[log in to unmask]>, +31 8886 64109  and Wessel Kraaij, [log in to unmask]<mailto:[log in to unmask]>, +31 8886 67194.

[Description: Description: cid:156293209@16032011-107B]<>

Prof. dr. ir. W. (Wessel) Kraaij
senior scientist
personal page<>

T +31 (0)88 866 71 94
M +31 (0)61 299 05 43
E [log in to unmask]<mailto:[log in to unmask]>



To unsubscribe from the MM-INTEREST list:
write to: mailto:[log in to unmask]
or click the following link: