JMLR Special Topic - Deadline extension

[Special Topic on Mining and Learning with Graphs]

Due to several requests and coincidence with ICML submission deadline, the deadline for the Special Topic of the Journal of Machine Learning Research on "Mining and Learning with Graphs and Relations" has been extended. Please see the updated call for papers below.

Mining and Learning with Graphs and Relations

a Special Topic of the Journal of Machine Learning Research

http://www.dsi.unifi.it/~paolo/JMLR08/

Call for Papers

As data mining and machine learning techniques continue to evolve and improve, the role  of structure in  the data becomes  more and more important.  A major driving force is the explosive growth in the amount of heterogeneous data that is being collected  in the business and scientific  world. Early approaches to statistical   learning   were   mainly   based  on   vector-based   data   and attribute-value  propositional representations. At  the end  of the  1990's, a "structured  revolution"  has started  to  profoundly  change  and extend  the representational  perspectives  in all  areas  of  machine  learning and  data mining. For  example, the widespread  diffusion of kernel methods  has allowed several  learning algorithms to  abstract away  data types  and be  applied to structured objects  simply by plugging-in  a suitable kernel function  for the data  type  at hand.  Yet,  research has  mainly  focused  on independent  and identically-distributed  (iid) examples.  Dealing with  inter-related examples that are linked together into complex graphs or hypergraphs remains one of the major  challenges. Similarly,  link  and relation  prediction, and  supervised learning  with structured  outputs are  substantially more  difficult problems than single-output classification or regression.

Dealing with  structured data has  deep unresolved foundational  and practical implications and affects different learning and mining paradigms. We therefore invite submission  from research communities working  on different theoretical and applicative aspects of machine  learning and data mining, especially those that are  active in cutting-edge frontier  topics. These include,  but are not limited to:

• (statistical) relational learning;
• (probabilistic) inductive logic programming;
• relational reinforcement learning;
• kernel methods for structured data;
• graph pattern discovery;
• subgraph mining;
• supervised learning with structured outputs  and/or collective predictions;
• multi-task and transfer learning;
• multi-relational data mining.

Application areas of interest are also diverse and include:

• web mining;
• bioinformatics;
• social networks analysis;
• information retrieval;
• natural language;
• chemoinformatics;
• robotics;
• communication networks;
• transportation networks.

Submission procedure:

A title and abstract must be sent by February 25, 2008 to mlgr...@dsi.unifi.it. The full manuscript must be submitted by March 3, 2008 using the JMLR submission system. Please follow the general JMLR author information when preparing your manuscript.

Important Dates:

New deadline for title and abstracts: February 25, 2008.

New submission deadline: March 3, 2008.

Notification to authors: May 12, 2008.

Revised manuscripts: July 7, 2008.

Guest Editors:

Paolo Frasconi, Università degli Studi di Firenze, Italy;

Kristian Kersting, CSAIL, MIT, Cambridge, USA;

Hannu Toivonen, University of Helsinki, Finland;

Koji Tsuda, Max Planck Institute for Biological Cybernetics, Tuebingen, Germany.

Guest editors can be contacted at mlgr...@dsi.unifi.it