====================================================================
CALL FOR PAPERS
====================================================================
WORKSHOP ON TRUST AND SECURITY IN ADAPTIVE RECOMMENDER SYSTEMS
June 20, 2006 - Dublin, Ireland
http://ihl.ucd.ie/ah2006/
at
4th International Conference on Adaptive Hypermedia and
Adaptive Web-Based Systems
June 21-23, 2006 - Dublin, Ireland
http://www.ah2006.org/
Background and Goals
====================
Recommender systems are a key component of many e-commerce and other
Internet-based applications. These systems assist users to find
relevant products and information from the ever-increasing quantities
that are available. Recommender systems have many known commercial
benefits, including turning web browsers into buyers, cross-selling
products, and instilling customer loyalty. However, these commercial
benefits come with risks, including misleading recommendations that
lead to unsatisfactory purchases, fake ratings or reviews by
interested parties, and misuse of personal data by on-line stores.
How can the value provided by a recommender system be evaluated
against the risk it creates?
New approaches are needed for evaluating recommender
systems. Previously, these systems have been evaluated primarily
across the dimensions of recommendation and predictive accuracy,
algorithm efficiency and scalability. As recommender systems become
more pervasive, however, new dimensions become increasingly important,
such as the honesty and reputation of users, the privacy of user
information, and the quality of user feedback.
The issues of trust and security in recommender systems have been
noted in several media publications. For example, in an article
entitled "United we find" (The Economist, March 10, 2005), concerns
were raised relating to the privacy of users' data and the possibility
that the output of recommendation systems can be manipulated. A New
York Times article of May 22, 2003 - "False Web Ratings Swing Opinion,
Study Says" - further discusses the issue of interfering with
recommender systems' recommendations. In addition, it has not been
unknown for clearly unrelated products to be surreptitiously linked in
certain cases. Such an example was reported on CNET News.com (December
6, 2002), where material of an extreme nature was surreptitiously
linked on Amazon.com with a spiritual guide authored by a well-known
televangelist. Such deliberate attacks and pranks pose a substantial
threat to the successful operation of the recommender system since any
behaviour that results in poor quality output is likely to undermine
the trust and confidence of system users.
Thus, there exists a need for robust recommendation algorithms that
are capable of dealing with malicious behaviour on the part of system
users, and that can offer users promises about the ways in which their
personal data may be used. The goal of this workshop is to bring
together academic researchers in the field of adaptive personalisation
in order to realise innovative solutions to recommendation that
satisfy the above considerations with respect to system integrity and
user privacy/security concerns.
Target Audience and Questions Addressed
=======================================
This workshop is relevant to all researchers in web
personalisation. The topic is important to anyone involved in the
development of personalisation systems that depend on databases of
user profile information. The focus is on:
* How the context in which such databases are created and grown
affects the security and privacy of the user;
* How the insertion of inaccurate information (both maliciously and
accidentally) impacts on the security and performance of the system;
* How systems can be developed to be robust to such inaccurate data,
for example by developing ways to identify the trustworthiness of
their user population, while still maintaining acceptable levels of
accuracy and coverage.
Workshop Topics
===============
The issues of trust and security in adaptive recommender systems and
web personalisation include a wide variety of topics. The following
list is not intended to be an exhaustive list for the workshop, and it
is only indicative of different research aspects and open problems:
* User modeling and profiling
* Trust and reputation models
* Influence models
* Privacy-preserving recommendations
* Social and privacy considerations
* Malicious attacks: detection and response
* Development of robust adaptive recommender systems
* Performance metrics related to trust
Workshop Format
===============
The workshop will run for one half day. Participation will be on the
basis of full papers (8-10 pages long). Authors of accepted papers
will give a 15 minute presentation followed by a 5 minute question and
answer session. The number of attendees to the workshop will be
limited in order to encourage participation.
Submissions in LNCS format only please (as per main conference). Go to:
http://www.springer.com/sgw/cda/frontpage/0,11855,5-164-2-72376-0,00.html#anchor\
16
for author instructions; an example PDF file can be downloaded from:
http://www.springer.com/sgw/cda/pageitems/document/cda_downloaddocument/0,11855,\
0-0-45-72797-0,00.pdf.
Please submit papers by email to the following address:
ahworkshop@..., to be received no later than March 31,2006. Due
to organisational constraints, late submissions cannot be considered.
Expected Results
================
We aim, through this workshop, to further promote awareness of the
importance of security and trust in the development of recommender
systems and personalisation in general. We hope that this workshop
will deliver new insights and techniques in these key areas and
provide a forum to allow researchers in this relatively-recently
emerged topic to come together and agree a framework for future
research in the field. Assuming that the workshop generates sufficient
interest, we intend as a next step to seek to develop the papers
presented at the workshop into a special issue of a relevant journal.
Workshop Co-Chairs
==================
Robin Burke
DePaul University, USA
Neil Hurley
University College Dublin, Ireland
Workshop Organisers
===================
Guenole Silvestre
University College Dublin, Ireland
Neil Hurley
University College Dublin, Ireland
Michael O'Mahony
University College Dublin, Ireland
Program Committee
=================
Robin Burke
DePaul University, USA
Neil Hurley
University College Dublin, Ireland
Nick Kushmerick
University College Dublin, Ireland
Bamshad Mobasher
DePaul University, USA
Michael O'Mahony
University College Dublin, Ireland
John Riedl
University of Minnesota, USA
Guenole Silvestre
University College Dublin, Ireland
Barry Smyth
University College Dublin, Ireland
Important Dates
===============
March 31: Submission deadline
May 5: Notification of acceptance
May 19: Final camera-ready copy
June 20: Workshop
Contact Information
===================
Robin Burke
School of Computer Science, Telecommunications and Information Systems,
243 S. Wabash Ave. #453l,
Chicago, IL 60604, USA
Phone: +1 312 362 5910
http://josquin.cti.depaul.edu/~rburke/
Neil Hurley
UCD School of Computer Science and Informatics,
University College Dublin,
Belfield, Dublin 4, Ireland
Phone: +353 1 716 2903
http://ihl.ucd.ie/people.php
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
Offline 
Send Email