Research Papers

ILASP has been developed as part of a comprehensive programme of research on the theory and practice of learning ASP programs. Mark Law's PhD thesis presents each of the ILASP algorithms in detail, along with theoretical results on learning ASP programs. Parts of the work in the thesis led to several conference and journal publications.

Thesis

[1]	Mark Law. Inductive learning of answer set programs. Imperial College London, 2018. [  pdf  ]

Selected Publications

[1]	Mark Law, Alessandra Russo, and Krysia Broda. Inductive learning of answer set programs. In Logics in Artificial Intelligence - 14th European Conference, JELIA 2014, Funchal, Madeira, Portugal, September 24-26, 2014. Proceedings, pages 311-325, 2014. [  paper | pdf | proofs  ]

[2]	Mark Law, Alessandra Russo, and Krysia Broda. Learning Weak Constraints in Answer Set Programming. In Theory and Practice of Logic Programming (TPLP), Proceedings of ICLP 15, 2015. [  pdf | proofs  ]

[3]	Mark Law, Alessandra Russo, and Krysia Broda. Iterative Learning of Answer Set Programs from Context Dependent Examples. In Theory and Practice of Logic Programming (TPLP), Proceedings of ICLP 16, 2016. [  pdf  ]

[4]	Mark Law, Alessandra Russo, and Krysia Broda. The complexity and generality of learning answer set programs. Artificial Intelligence, 2018. [  pdf  ]

[5]	Mark Law, Alessandra Russo, and Krysia Broda. Inductive Learning of Answer Set Programs from Noisy Examples. In Advances in Cognitive Systems, 2018. [  pdf  ]

[6]	Mark Law, Alessandra Russo, and Krysia Broda. The ILASP system for Inductive Learning of Answer Set Programs. To appear in The Association for Logic Programming Newsletter, 2020. [  pdf  ]

Information about tasks used as benchmarks for ILASP

The rest of this page provides learning tasks which have been used to evaluate ILASP in our research.

• Details of Hypothesis Spaces from the paper "Inductive Learning of Answer Set Programs from Noisy Examples"

  Details of the hypothesis spaces used in the above paper are given in this supplementary document.

• Tasks from the paper "Iterative Learning of Answer Set Programs from Context Dependent Examples"

  We give here encodings of the tasks which were run for the paper "Iterative Learning of Answer Set Programs from Context Dependent Examples", which was presented at ICLP 2016 and published in TPLP. To replicate the experiments in the paper, all tasks should be run with ILASP v2.6.0. The encodings are explained in greater detail in this document.

  • Hamilton A
  • Hamilton B
  • Scheduling A
  • Scheduling B
  • Scheduling C
  • Agent A
  • Agent B
  • Agent C
  • Agent D
  • Journey

• Noisy Hamilton learning tasks

  We give here encodings of the tasks which we use to evaluate the ILASP3 algorithm. All tasks are noisy versions of the Hamilton learning setting, with 5% noise and varying numbers of examples.

  • 20 examples
  • 40 examples
  • 60 examples
  • 80 examples
  • 100 examples
  • 120 examples
  • 140 examples
  • 160 examples
  • 180 examples
  • 200 examples

• Tasks from the paper "The ILASP system for Inductive Learning of Answer Set Programs"

  We give here encodings of the tasks referred to in the paper "The ILASP system for Inductive Learning of Answer Set Programs".

  • coin1.las
  • coin2.las
  • coin3.las
  • hamilton.las
  • journey.las