Machine Learning Journal

Special Issue on

Inductive Logic Programming and Relational Learning

Stan Matwin and Claude Sammut, Guest Editors

Deadline extended - see the end of this CFP


We are pleased to announce that the Machine Learning Journal will again publish a special issue on Inductive Logic Programming and Relational Learning, following the 13th International Conference on Inductive Logic Programming held in Sydney, Australia, in July 2002.



The Editors encourage submissions to the Special Issue of MLJ papers on topics  which are generally considered within the scope of relational learning and inductive logic programming. The list of topics below is indicative of the range of the special issue, but it is by no means complete:


     - first-order rule learning

     - learning probabilistic and statistical relational models

     - link discovery

     - learning from relational databases

     - learning in representations other than logic (e.g. learning tree

       patterns, learning conceptual graphs)

     - (multi)relational data mining

     - ILP-specific issues in relational data mining outside model building

       (sampling, evaluation, etc.)

     - multi-instance learning

     - change of representation and propositionalization

     - relational reinforcement learning

     - instance-based and kernel-based methods in ILP

     - relational text and Web mining; learning from semi-structured data

       (e.g. XML)

     - learnability and algorithmic issues of learning from structured data

     - inductive databases

     - learning in higher-order logics

     - logical foundations of induction

     - efficient implementations

     - practical applications of ILP and Relational Data Mining


We particularly encourage papers which  combine ILP with methods and tools from other areas of Machine Learning or AI in general, e.g. learning probabilistic concept descriptions, or ILP in Natural Language Processing. We also encourage papers in which new, interesting applications are addressed by means of relational/ILP methods.




Authors of papers that appeared previously in refereed conferences and workshops (e.g. in ILP 2002) are encouraged to submit extended version of these papers. Such papers must contain material that is new wrt the conference paper, e.g. to provide more details, to give greater, more in-depth, discussion of the results and related work, to expand upon the experimental results, and to give a more thorough and scholarly treatment of the material than is possible in a conference paper. Moreover, all papers have to be accessible to the broad readership of the journal, not just to researchers in ILP and relational learning. Submissions must not have appeared in, nor be under consideration by, other journals. Authors of papers whose previous versions appeared in refereed conferences and workshops are requested to include in their submission a brief letter stating the differences between the previously published version of their paper and this MLJ Special Issue submission.




   Manuscripts should be prepared for 8 1/2 x 11 in paper, with pages numbered consecutively. Papers should be at most 20,000 words in length, with full-page figures counting for 400 words. Detailed guidelines are at:


   We kindly ask that all submissions be made electronically, as a postscript or pdf attachment to:


   Since this is a special issue, please be sure that:


-          the topic of the email containing your submission must be Submission to the Special Issue of  MLJ  on ILP/RL

-          your submission is cc’ed to:, for double-checking


   If you cannot submit your paper electronically, please send five (5) hardcopies of the manuscript to Kluwer at the address specified in the instructions, and one (1) hardcopy to:


Stan Matwin

SITE, 800 King Edward

University of Ottawa

Ottawa, ON

K1N 6N5



With a note on the envelope: MLJ ILP-RL submission.



All inquiries regarding this special issue should be directed to Stan Matwin (





Submission deadline:           March  20, 2003