GPU-Accelerated Hypothesis Cover Set Testing for Learning in Logic

Eyad Algahtani; Dimitar Lubomirov Kazakov

GPU-Accelerated Hypothesis Cover Set Testing for Learning in Logic

Eyad Algahtani, Dimitar Lubomirov Kazakov

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

ILP learners are commonly implemented to consider sequentially each training example for each of the hypotheses tested. Computing the cover set of a hypothesis in this way is costly, and introduces a major bottleneck in the learning process. This computation can be implemented more efficiently through the use of data level parallelism. Here we propose a GPU-accelerated approach to this task for propositional logic and for a subset of first order logic. This approach can be used with one’s strategy of choice for the exploration of the hypothesis space. At present, the hypothesis language is limited to logic formulae using unary and binary predicates, such as those covered by certain types of description logic. The approach is tested on a commodity GPU and datasets of up to 200 million training examples, achieving run times of below 30ms per cover set computation.

Original language	English
Title of host publication	CEUR Proceedings of the 28th International Conference on Inductive Logic Programming
Publisher	CEUR Workshop Proceedings
Publication status	E-pub ahead of print - 2018

Access to Document

ilp2018-algahtani-kazakov-finalAccepted author manuscript, 1.33 MBLicence: GNU GPL

Cite this

@inproceedings{2598f5c61e464891bd20b1b05176c3f4,

title = "GPU-Accelerated Hypothesis Cover Set Testing for Learning in Logic",

abstract = "ILP learners are commonly implemented to consider sequentially each training example for each of the hypotheses tested. Computing the cover set of a hypothesis in this way is costly, and introduces a major bottleneck in the learning process. This computation can be implemented more efficiently through the use of data level parallelism. Here we propose a GPU-accelerated approach to this task for propositional logic and for a subset of first order logic. This approach can be used with one{\textquoteright}s strategy of choice for the exploration of the hypothesis space. At present, the hypothesis language is limited to logic formulae using unary and binary predicates, such as those covered by certain types of description logic. The approach is tested on a commodity GPU and datasets of up to 200 million training examples, achieving run times of below 30ms per cover set computation.",

author = "Eyad Algahtani and Kazakov, {Dimitar Lubomirov}",

year = "2018",

language = "English",

booktitle = "CEUR Proceedings of the 28th International Conference on Inductive Logic Programming",

publisher = "CEUR Workshop Proceedings",

}

TY - GEN

T1 - GPU-Accelerated Hypothesis Cover Set Testing for Learning in Logic

AU - Algahtani, Eyad

AU - Kazakov, Dimitar Lubomirov

PY - 2018

Y1 - 2018

N2 - ILP learners are commonly implemented to consider sequentially each training example for each of the hypotheses tested. Computing the cover set of a hypothesis in this way is costly, and introduces a major bottleneck in the learning process. This computation can be implemented more efficiently through the use of data level parallelism. Here we propose a GPU-accelerated approach to this task for propositional logic and for a subset of first order logic. This approach can be used with one’s strategy of choice for the exploration of the hypothesis space. At present, the hypothesis language is limited to logic formulae using unary and binary predicates, such as those covered by certain types of description logic. The approach is tested on a commodity GPU and datasets of up to 200 million training examples, achieving run times of below 30ms per cover set computation.

AB - ILP learners are commonly implemented to consider sequentially each training example for each of the hypotheses tested. Computing the cover set of a hypothesis in this way is costly, and introduces a major bottleneck in the learning process. This computation can be implemented more efficiently through the use of data level parallelism. Here we propose a GPU-accelerated approach to this task for propositional logic and for a subset of first order logic. This approach can be used with one’s strategy of choice for the exploration of the hypothesis space. At present, the hypothesis language is limited to logic formulae using unary and binary predicates, such as those covered by certain types of description logic. The approach is tested on a commodity GPU and datasets of up to 200 million training examples, achieving run times of below 30ms per cover set computation.

M3 - Conference contribution

BT - CEUR Proceedings of the 28th International Conference on Inductive Logic Programming

PB - CEUR Workshop Proceedings

ER -