v4.4.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features/changes in v4.4.0

1. Minor bug fixes.
2. Experimental support for learning heuristics
3. Overhaul internal conflict analysis implementation
4. --tiny-constraints flag.

v4.3.1

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features/changes in v4.3.1

1. Minor bug fixes.

v4.3.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features/changes in v4.3.0

1. Minor bug fixes.
2. Support for changing clingo arguments through PyLASP.
3. Ordering examples between an answer set and a fixed cost vector (rather than two answer sets) -- more details to follow.

v4.2.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features/changes in v4.2.0

1. Minor bug fixes.
2. Additions to the PyLASP API. ILASP2 is now fully integrated with PyLASP, meaning that violating reasons etc can be displayed to the user for the first time.
3. ILASP is now built against static Clingo libraries, meaning that it is no longer necessary to install a particular Clingo version when using ILASP.
4. We have dropped Lua support for this release. We are hoping to reintroduce it in a future release.

v4.1.2

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features in v4.1.2

1. Minor bug fixes.
2. Experimental new type or ordering example (details to follow)

v4.1.1

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features in v4.1.1

1. Support for Clingo 5

2. Partial support for meta-injection in ILASP 2i.

v4.1.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features in v4.1.0

1. Support for ASP absolute value function.

2. Support for aggregates in the body of background/context rules.

v4.0.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features in v4.0.0

1. A new algorithm for solving ILASP tasks (run using --version=4). Both ILASP3 and ILASP4 use a conflict-driven approach to ILP, computing constraints which must hold for example to be covered. ILASP3 computes very large constraints which are guaranteed to be both necessary and sufficient for an example to be covered (which is computationally expensive). ILASP4 computes constraints which are only guaranteed to be necessary. This is often much cheaper (computationally).

2. PyLASP scripts. This is a new feature allowing users much more control over ILASP. It is now possible to tell ILASP not to use any of the built-in algorithms, and users can instead define their own algorithms in Python using modular parts of ILASP. ILASP 2i, 3 and 4 have all been (re)implemented as PyLASP scripts. Details of the PyLASP built-in functions are available here. To see the PyLASP implementation for ILASP, run ILASP on any learning task with a set of configuration options for the PyLASP script to use (e.g. --version=3 will generate a different PyLASP script to --version=4), together with the new flag "-p". This will cause ILASP to output a PyLASP script, which can then be customised.

More details and documentation on these two features will be made public in the next month.

New dependencies

1. Due to the new PyLASP features, ILASP now depends on Python 2.7.
2. ILASP now calls Clingo 5 using its C API. From this version, to run ILASP, you will need libclingo (version 3) to be installed. In most cases, when interacting with Clingo, ILASP uses this API, but there are some cases where it is faster to call the executable, so Clingo 5 should still be in the PATH as usual.

v3.6.0

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, disjunctive rules, choice rules, constraints, and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If you do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

New Features in v3.6.0

1. Conditional literals. These can be used just as they are in Clingo, and can be used in the background knowledge, example contexts, and the hypothesis space. The predicates which can be used as conditions in a conditional literal can be declared using the new #modec declaration.

2. Disjunction. Again this can be used in the background knowledge, example contexts, and the hypothesis space. The syntax of disjunctive rules is just the same as in Clingo. To enable learning disjunctive rules, the user should add the line #bias("allow_disjunction.").

For more details on these new features, please see the new ILASP manual.

Minor bug-fixes

1. Fix issue which stopped empty contexts working in the same task as non-empty contexts.
2. Other very minor bug fixes.

v3.5.1

ILASP

ILASP (Inductive Learning of Answer Set Programs) is a new logic-based learning system capable of learning normal rules, choice rules, constraints and weak constraints in ASP.

Please note that the source code links on this page have been auto-generated by GitHub, and do not contain the ILASP source.

Usage

ILASP is free to use for non-commercial research and education. If do use ILASP for research, we
ask that you use this citation. Anyone wishing to use ILASP for commercial purposes should contact Mark Law (mark@ilasp.com). For details and examples of how to use ILASP, please see the manual (which is available at www.ilasp.com).

Bug Reports and Feature Requests

Please submit all bug reports and feature requests as issues on this GitHub repository.

Minor bug-fixes

1. Fix issue with --strict-types flag causing some rules with negation as failure to be omitted from the hypothesis space.

2. Do not sort the hypothesis space if it has been explicitly specified (doing so can break meta-injected specifications).