Generalisation through negation and predicate invention
The ability to generalise from a small number of examples is a fundamental challenge in machine learning. To tackle this challenge, we introduce an inductive logic programming (ILP) approach that combines negation and predicate invention. Combining these two features allows an ILP system to generali...
| Հիմնական հեղինակներ: | , |
|---|---|
| Ձևաչափ: | Conference item |
| Լեզու: | English |
| Հրապարակվել է: |
Association for the Advancement of Artificial Intelligence
2024
|
| _version_ | 1826313018485506048 |
|---|---|
| author | Cerna, DM Cropper, A |
| author_facet | Cerna, DM Cropper, A |
| author_sort | Cerna, DM |
| collection | OXFORD |
| description | The ability to generalise from a small number of examples is a fundamental challenge in machine learning. To tackle this challenge, we introduce an inductive logic programming (ILP) approach that combines negation and predicate invention. Combining these two features allows an ILP system to generalise better by learning rules with universally quantified body-only variables. We implement our idea in NOPI, which can learn normal logic programs with predicate invention, including Datalog programs with stratified negation. Our experimental results on multiple domains show that our approach can improve predictive accuracies and learning times. |
| first_indexed | 2024-03-07T08:23:53Z |
| format | Conference item |
| id | oxford-uuid:e63e420c-976b-4d1c-b24d-f91875b4d350 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-09-25T04:04:19Z |
| publishDate | 2024 |
| publisher | Association for the Advancement of Artificial Intelligence |
| record_format | dspace |
| spelling | oxford-uuid:e63e420c-976b-4d1c-b24d-f91875b4d3502024-05-13T12:10:56ZGeneralisation through negation and predicate inventionConference itemhttp://purl.org/coar/resource_type/c_5794uuid:e63e420c-976b-4d1c-b24d-f91875b4d350EnglishSymplectic ElementsAssociation for the Advancement of Artificial Intelligence2024Cerna, DMCropper, AThe ability to generalise from a small number of examples is a fundamental challenge in machine learning. To tackle this challenge, we introduce an inductive logic programming (ILP) approach that combines negation and predicate invention. Combining these two features allows an ILP system to generalise better by learning rules with universally quantified body-only variables. We implement our idea in NOPI, which can learn normal logic programs with predicate invention, including Datalog programs with stratified negation. Our experimental results on multiple domains show that our approach can improve predictive accuracies and learning times. |
| spellingShingle | Cerna, DM Cropper, A Generalisation through negation and predicate invention |
| title | Generalisation through negation and predicate invention |
| title_full | Generalisation through negation and predicate invention |
| title_fullStr | Generalisation through negation and predicate invention |
| title_full_unstemmed | Generalisation through negation and predicate invention |
| title_short | Generalisation through negation and predicate invention |
| title_sort | generalisation through negation and predicate invention |
| work_keys_str_mv | AT cernadm generalisationthroughnegationandpredicateinvention AT croppera generalisationthroughnegationandpredicateinvention |