A Framework for Learning Rules from Multiple Instance Data

  • Yann Chevaleyre
  • Jean-Daniel Zucker
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2167)


This paper proposes a generic extension to propositional rule learners to handle multiple-instance data. In a multiple-instance representation, each learning example is represented by a bag” of fixed-length feature vectors”.Such a representation,lying somewhere between propositional and first-order representation, offers a tradeoff between the two. Naive-RipperMi is one implementation of this extension on the rule learning algorithm Ripper. Several pitfalls encountered by this naive extension during induction are explained. A new multiple-instance search bias based on decision tree techniques is then used to avoid these pitfalls. Experimental results show the benefits of this approach for solving propositionalized relational problems in terms of speed and accuracy.


Multiple Instance Target Concept Inductive Logic Programming Candidate Rule Decision Tree Technique 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Erick Alphonse and Celine Rouveirol. Lazy propositionalization for relational learning. In ECAI, 2000.Google Scholar
  2. 2.
    P. Auer, P. Long, and Ashwin Srinivasan. Approximating hyper-rectangles: Learning and pseudo-random sets. In Annual ACM Symposium on Theory of Computing, 1997.Google Scholar
  3. 3.
    Peter Auer. On learning from multi-instance examples: Empirical evaluation of a theoretical approach. In Proc. 14th International Conference on Machine Learning, 1997.Google Scholar
  4. 4.
    Hendrik Blockeel, Luc De Raedt, Nico Jacobs, and Bart Demoen. Scaling up inductive logic programming by learning from interpretations. Data Mining and Knowledge Discovery, 3(1):59–93, 1999.CrossRefGoogle Scholar
  5. 5.
    Avrim Blum and Adam Kalai. A note on learning from multiple-instance examples. Machine Learning, 30, 1998.Google Scholar
  6. 6.
    Yann Chevaleyre and J.D. Zucker. Noise tolerant rule induction for multiple-instance data and potential data mining application. Tech. Rep. University of Paris 6, available at, 2001.
  7. 7.
    William W. Cohen. Fast effective rule induction. In Proc. 12th International Conference on Machine Learning. Morgan Kaufmann, 1995.Google Scholar
  8. 8.
    Thomas G. Dietterich, Richard H. Lathrop, and Tomás Lozano-Pérez. Solving the multiple instance problem with axis-parallel rectangles. Artificial Intelligence, 89(1–2), 1997.Google Scholar
  9. 9.
    Eibe Frank and Ian H.Witten. Generating accurate rule sets without global optimization. In Proc. 15th ICML, 1998.Google Scholar
  10. 10.
    Attilio Giordana, Lorenza Saitta, Michele Sebag, and Marco Botta. Analyzing relational learning in the phase transition framework. In Proc. 17th ICML, 2000.Google Scholar
  11. 11.
    J.U. Kietz. Some lower bounds for the computational complexity of inductive logic programming.In ECML, 1993.Google Scholar
  12. 12.
    Oded Maron and Aparna Lakshmi Ratan. Multiple-instance learning for natural scene classification.In Proc. 15th ICML, pages 341–349, 1998.Google Scholar
  13. 13.
    Luc De Raedt. Attribute-value learning versus inductive logic programming: The missing links. In Proc. 8th International Conference on ILP, 1998.Google Scholar
  14. 14.
    Michele Sebag and Celine Rouveirol. Tractable induction and classification in first order logic. In IJCAI, Nagoya, Japan, 1997.Google Scholar
  15. 15.
    A. Srinivasan and S. Muggleton. Comparing the use of background knowledge by two ilp systems. In L. de Raedt, editor, ILP-95., Katholieke Universiteit Leuven, 1995.Google Scholar
  16. 16.
    Jun Wang and Jean-Daniel Zucker. Solving multiple-instance problem: a lazy learning approach.In Proc. 17th ICML, 2000.Google Scholar
  17. 17.
    Jean-Daniel Zucker and Jean-Gabriel Ganascia. Learning structurally indeterminate clauses.In Proc. 8th International Conference on ILP. Springer-Verlag, 1998.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Yann Chevaleyre
    • 1
  • Jean-Daniel Zucker
    • 1
  1. 1.LIP6-CNRSUniversity Paris VIParis Cedex 05France

Personalised recommendations