Find the most recent information on EU Funding activities in the field of Information and Communication Technologies (ICT) by visiting our ICT in FP7 website, which covers ICT in the 7th Framework Programme (FP7) 2007 - 2013.
PENG - Personalized News Content Programming
The PENG project aimed at providing news professionals with an interactive and personalized tool for multimedia news gathering and delivery. This was achieved by developing a flexible prototype for a personalized filtering, retrieval and composition of multimedia news. PENG aims at yielding personalized multimedia news customized for distinct users, based on the analysis of users' profiles encoding users' preferences and interests allowing an effective news fruition.
Impact
PENG looked to define a standard technological solution that supports heterogeneous users: it was supposed to affect a wide and very diverse community of users, with different needs and requirements. Possible user targets were considered to be journalists, information-intensive workers, students of communication faculties. The target user community where we were aiming to disseminate the PENG project was a very specific community with uneven IT skills.
PENG's main innovation
The main innovative aspects of the PENG system are the following:
- the filtering phase is based on an innovative multi-criteria decision making model. Among the considered criteria the user can "tune" the contribution of the distinct content types and sources of information by associating a trust score to each information source.
- the PENG system allows a fuzzy hierarchical categorization of news (based on a new algorithm defined within the PENG project) so as to deal with vagueness and imprecision in the filtering task when associating news to a category subject matter or topic area: news can belong to several categories and deal at the same time with different subject matters to a different strength and deepness. This fuzzy categorization avoids professionals to miss interesting news when selecting and browsing into a single category.
- the PENG system allows learning the user context: disambiguation using pre-learned (global or personal) knowledge or dynamically learned user preferences changing in time allows a more natural expression of the user's needs and makes the system adaptive to the user context (use of content based filtering techniques).
- the PENG system adopts innovative techniques of distributed Information Retrieval.
PENG's results
The PENG system has been designed with five main modules: user profile manager, filtering, retrieval, presentation and common database manager modules. The user profile manager and the common database manager module communicate with all the other modules. These modules reside on the PENG server. The other modules do not communicate directly but always via the user profile manager and the common database manager. The user accesses the system locally through a GUI.
During the first year of the project the requirements analysis and the system architecture have been defined and the structure of the main system modules has been designed.
Currently all the three single modules have been implemented, tested and evaluated from a strict technical point of view and now we are in the phase of their integration.
The requirement analysis phase aided in identifying the typical users of PENG, their expectations with respect to a new system helping them in carrying out their work.
The observation of journalists in performing their habitual working activity has also been essential to define the PENG typical scenarios and its use cases.
More details
- PENG Website
- Project Showcase (Video) (18.5MB AVI zipped) and Showcase Scenario (PDF, 18KB)
- IST Results article: Refining the process of gathering information (25 September 2006)(PDF, 134KB)
- Brochure (PDF, 417KB)
- Presentation (PDF, 432KB)
- Annual Report 2006 (PDF, 53KB)
- Annual Report 2005 (PDF, 234KB)
- Presentation of PENG Project at:
- the European Workshop on the Integration of Knowledge, Semantics and Digital Media Technology, in London on November 25-26, 2004
- Department of Computer Science of the Università degli Studi di Milano Bicocca
- Department of Computer Science of the Università degli Studi di Bergamo
- Department of Lugano University
- Glasgow University Information Retrieval group.
- Xerox PARC, Palo Alto, California, USA
- I-LAB GROUP in the Department of CIS in University of Strathclyde (Glasgow)
- IST Conference, The Hague
- the ATOS Marketing Department
- the CONTENT4ALL (Cross Platform Tools for Community Content Publishing) project group
- South East Europe TV (SEE TV), TV Andalucia through Global Sevilla organisation and European Initiative on Networked and Electronic Media (NEM) and New Media Council
- the Division Media department of ATOS Origin
- First International Conference on Collaborative Computing
- Telemadrid
- all media agencies who collaborate with Radio Televisione Svizzera Italiana (AFP, ANSA, AP, DPA, Interfax, Reuters, SDA/ATS, SI)
- Publications on innovative aspects of PENG:
- G.Pasi "An Overview of the PENG (Personalized News content programminG) Project" , in the proceedings of the European Workshop on the Integration of Knowledge, Semantics and Digital Media Technology, November 25-26, 2004, London, U.K
- G..Bordogna, M. Pagani, G.Pasi, "A dynamical Hierarchical fuzzy clustering algorithm for document filtering", accepted for publication in "Soft Computing for Information Retrieval on the Web", Springer Verlag, 2005
- G.Pasi, R.Villa, "Personalized News Content Programming (PENG): A System Architecture", in the proceedings of the International DEXA Workshop Integrating Data Mining, Databases and Information Retrieval (IDDI-05), Copenhagen, Denmark, August 22, 2005
- G..Bordogna, M. Pagani, G.Pas, L. Antoniolli, F. Invernizzii "An Incremental Hierachical Fuzzy Clustering Algorithm Supporting News Filtering" in the proceedings of the "Information Processing and Management of Uncertainty", Paris, July 2-7, 2006
- Papers:
- F. Crestani, M. Baillé "Analysing Measures of Resource Description Quality for Distributed Information Retrieval" (paper submitted to the 21st Annual ACM Symposium on Applied Computing)
- F. Crestani, M. Baillé "Towards Better Measures: Evaluation of estimated resource description quality for DIR" (paper submitted to the IEEE Infoscale Conference)
- F. Crestani, M. Baillé "When Enough is Enough: A Principled Termination Criterion for Query-Based Sampling" (paper submitted to the European Conference on Information Retrieval (ECIR))
- F. Crestani, M. Baillé "The University of Strathclyde at TREC HARD" (paper submitted for to the Text Retrieval Conference (TREC))
Administrative Details
- PENG started on 1 September 2004 and finished on 31 August 2006.
- The overall budget is roughly 1.8 million euro.
- Six partners from five countries have been involved in the project.
List of Participants
- CONSIGLIO NAZIONALE DELLE RICERCHE ITC-CNR
Via Bassini N. 15 20131 Milan, Italy - UNIVERSITA' DELLA SVIZZERA ITALIANA
Facolta' di Scienze della Comunicazione, via Buffi 13, 6904 Lugano, Switzerland - ATOS ORIGIN SOCIEDAD ANONIMA ESPANOLA
Albarracín 25, Madrid 28037, Spain - UNIVERSITY OF STRATHCLYDE
Richmond Street 16, Glasgow, G11XQ, UK - UNIVERSITE' JOSEPH FOURIER GRENOBLE 1
Avenue Centrale 621, Domaine Universitaire, Grenoble,Cedex 9,BP 53, 38042, France - SOCIETA' SVIZZERA DI RADIOTELEVISIONE
Via Canevascini 1, Lugano CH6903, 6900, Switzerland
Contact Persons
- Project Coordinator: Gabriella Pasi
- CONSIGLIO NAZIONALE DELLE RICERCHE ITC-CNR
- Via Bassini 15
- 20133 Milano, Italy
- Email: Gabriella Pasi
Events in connection with PENG
- IPMU 2006 - Information Processing and Management of Uncertainty in Knowledge-based Systems, 2-7 July, Paris
- IST 2006, 22-24 Nov, Helsinki
- eChallenges 2006, Barcelona
- Spire 2006 - 13th Edition of the Symposium on String Processing and Information Retrieval, 11-13 Oct, Glasgow