Inductive Logic Programming Through Worst-Case Optimal Join Optimization Based on Variable-Elimination Ordering and Bitmap Indexing
Year: 2026
Type: software
Abstract: Scala Inductive Logic Programming (SiLP) SiLP is an inductive logic programming (ILP) framework written entirely in Scala, combining functional and object-oriented programming paradigms. Its core functionalities include query optimization, query processing, unification, and predicate invention. SiLP uses Prolog-like syntax and is fully compatible with basic Prolog. It supports aggregate functions such as counting, computing distinct elements, and performing operations over lists. While these features are available within the framework, they have not been extensively tested. The system also supports recursive queries, meta-rule-based predicate invention, early stopping, pruning strategies during predicate search, and parallel processing. Although the current version does not support probabilistic reasoning, future versions are planned to include probabilistic extensions over the predicate space. SiLP is conceptually similar to Popper, though not as advanced, as its functionality is limited to pure Scala code and does not integrate external Prolog programs. Its main advantage lies in being a user-friendly, open-source library that does not require a separate platform for query processing. Files The directory structures contains documentation, source code, experiments directory, and examples for Prolog databases without macros and function definitions. 1. **/examples/** Each example is a separate sub-directory. It contains a query file, database facts, positive/negative samples in query.pl, bk.pl, and exp.pl respectively. Exp.pl is used for predicate invention evaluation through Experiments class.2. **/experiments/** The experiments include the output files for database statistics, and performance evaluations in statistics.csv, and performance.csv file respectively. The statistics file is generated with-in *DatabaseStatictics* class, and performance file is generated with-in *Performance* class.3. **/docs/** Includes this readme file and contains code examples for different categories. Documentation There is not installation setup for the source code, it can be important through maven supported IDE. Maven files contains required Scala dependencies. Scala 3.0 and Java-11 is required to run the project. 1. [Data structures and functionality](https://github.com/volkanagun/ILPEngine/blob/master/docs/Data%20structures.md) : This is a general overview that contains code snippets without going further details. The main classes required for query processing and optimization, indexing, unification and answer set programming, predicate structures, and predicate invention.2. [Unification and predicates](https://github.com/volkanagun/ILPEngine/blob/master/docs/Unification.md) : Unification examples and a guide for defining new predicates.3. [Query Optimization](https://github.com/volkanagun/ILPEngine/blob/master/docs/Query%20Optimization.md) : The documentation contains query optimization examples with a general overview of different optimization algorithms.4. [Predicate Invention](https://github.com/volkanagun/ILPEngine/blob/master/docs/Predicate%20Invention.md): It contains detailed examples including how to tune the parameters for predicate invention. Support Please refer the articles mentioning the SiLP in Google Scholar for supporting this project. Also you can give a link for the Github project named ILPEngine.
Author: Agun, Hayri Volkan
Institution: Bursa Technical University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7163045064