Temporal Datalog with Existential Quantification

Authors: Matthias Lanzinger, Markus Nissl, Emanuel Sallinger, Przemysław A. Wałęga

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we report first experiments with our prototypical implementation of Datalog MTL with uniform semantics and OWA, to demonstrate that our formalism has a potential for practical feasibility.
Researcher Affiliation Academia 1 Department of Computer Science, University of Oxford 2 TU Wien matthias.lanzinger@cs.ox.ac.uk, markus.nissl@tuwien.ac.at, sallinger@dbai.tuwien.ac.at, przemyslaw.walega@cs.ox.ac.uk
Pseudocode No The paper describes the implementation approach textually but does not include any pseudocode or algorithm blocks.
Open Source Code No Since Vadalog is a commercial system we are not able to provide a full access to our implementation. Instead, we provide an online tool (https://kg.dbai.tuwien.ac.at/vadalog-scheduler), which allows to pass Datalog MTL instances to our implementation, as well as to view the obtained outputs.
Open Datasets Yes Our first benchmark is based on data from Eclipse Simulation of Urban MObility (SUMO)1 describing road vehicles in a traffic jam, which was used during the Hackathon Challenge at the Stream Reasoning Workshop 2021 [Schneider et al., 2022]. As in the Hackathon, we considered a simple map of roads and three dataset D1, D2, and D3 corresponding to small, medium, and large levels of traffic, respectively. The datasets have roughly 800, 5,000, and 9,000 temporal facts describing vehicles position, speed, acceleration, and direction, among others. 1http://www.eclipse.org/sumo/ Our second benchmark exploits i Temporal, which allows us to generate Datalog MTL programs and matching datasets of varying size [Bellomarini et al., 2022b]. ... for each Πi, we invoke i Temporal to generate three matching datasets that contain approximately 5000, 50,000, and 500,000 temporal facts, respectively.
Dataset Splits No The paper describes the datasets used and their generation but does not provide specific details on how these datasets were split into training, validation, or test sets.
Hardware Specification Yes All our experiments were run on an Intel Core i7-8700 CPU with 64GB memory.
Software Dependencies No The paper mentions "Vadalog system" and tools like "Eclipse Simulation of Urban MObility (SUMO)" and "i Temporal" but does not provide specific version numbers for these software components.
Experiment Setup No The paper describes the Datalog MTL programs (rules) used in the experiments and the general approach (Skolemisation, chase procedure) but does not provide specific experimental setup details such as hyperparameters or system-level configuration settings for running the experiments.