CfP: ACM TOIT Special Issue on Machine Learning for the Internet

[Marco Maggini]

------------------------------------------------------
We apologize, if you receive multiple copies.
Please feel free to publicize.
Thank you.
------------------------------------------------------

                      CALL FOR PAPERS
           ACM Transactions on Internet Technology

                     Special Issue on
              Machine Learning for the Internet



Machine learning methods are becoming increasingly important for the 
development of several internet related technologies. Tasks such as 
intelligent searching, organizing, retrieving, and filtering information 
on the Web are extremely challenging and still much too easy for humans 
than they are for computers, except that humans are unable to scale up 
with the enormous amount of available data. Explicit coding of rules in 
this domain is typically very hard, and even if possible, would require 
exceptional coordination efforts. In particular, the fast dynamics of 
the information available on the Internet requires new approaches for 
indexing. The organization of information in Internet portals is 
becoming hardly manageable by humans. The users' surfing of the Internet 
can be made easier by personalized tools like search engines optimized 
for a specific Web community or even for the single user. For example, 
finding relevant documents by querying a search engine with a set of 
keywords may be difficult unless a proper ranking scheme is used to 
order the results. In this case, techniques based on user profiles, on 
topic selection and on the use of the Web topology can help in defining 
authoritative sources of information with respect to the given query and 
interests.

Searching, organizing and retrieving information from the Web poses new 
issues that can be effectively tackled by applying machine learning 
techniques. Learning algorithms can be used to devise new tools which 
improve the accessibility to the information available on the Web. 
Learning is particularly useful to automate those tasks in which it is 
quite easy to collect examples while coding a set of explicit rules is 
impractical. For example, the fast dynamics of the Internet can be faced 
by designing new specialized search tools which cover only the parts of 
the Web related to a given topic. These search tools focus their 
exploration only on the portion of the Web which contains the 
information relevant for this topic. Moreover, learning-based search 
tools can feature a very high precision in retrieving information and 
can reduce the need for human efforts for many repetitive tasks (like 
organizing documents in Web directories).

Beside accessing information, understanding and characterizing web 
structure and usage is essential for future development and organization 
of new tools and services. In spite of several recent efforts in 
measuring and producing mathematical models of web connectivity, 
dynamics, and usage, no definitive answers have emerged and learning may 
play a fundamental role for advancing our understanding in this field.

Papers are invited on applications of machine learning to all aspects of 
Internet technology. These include (but are not limited to):

     * Automated creation of web directories
     * Automatic extraction of information from Web pages
     * Automatic security management
     * Categorization of web pages
     * Design and improvement of web servers through prediction of
       request patterns
     * Focused crawling
     * Information retrieval for the design of thematic search engines
     * Models and laws that characterize the web structure
     * User modeling for the personalization of Web services

Submissions

Authors are requested to send an intention of submission (with authors, 
title and abstract) as an email message in plain text to 
acm-toit at dsi.unifi.it by May 1, 2002. Then, papers must be submitted in 
electronic format as an attachment to the same email address before May 
15, 2002. Preferred formats are PDF and PostScript (compressed with gzip 
or zip). Manuscripts must not exceed 50 single-column, double-spaced 
pages (including figures and tables) and must be written in English and 
set in 10 or 11 point font. Please do not send papers directly to guest 
editors' email addresses.

Important Dates

Intention of submission: May 1, 2002
Submission deadline:  May 15, 2002
Notification:  August 1, 2002

Guest editors

Gary William Flake
NEC Research Institute
4 Independence Way
Princeton, NJ 08540 (USA)
flake at research.nj.nec.com
Voice: +1 609-951-2795
http://www.neci.nj.nec.com/homepages/flake/

Paolo Frasconi
Dept. of Systems and Computer Science
University of Florence
Via di Santa Marta 3, I-50139 Firenze (Italy)
paolo at dsi.unifi.it
Voice: +39 055 4796 362
http://www.dsi.unifi.it/~paolo/

C. Lee Giles
School of Information Sciences and Technology
The Pennsylvania State University
001 Thomas Building,
University Park, PA, 16802 (USA)
giles at ist.psu.edu
Voice: +1 814 865 7884
http://ist.psu.edu/giles/

Marco Maggini
Dept. of Information Engineering
University of Siena
Via Roma 56, I-53100 Siena (Italy)
maggini at dii.unisi.it
Voice: +39 0577 233696
http://www.dii.unisi.it/~maggini/