Constraint acquisition and the data collection bottleneck

Thumbnail Image
Files
Date
2022-02-28
Authors
Prestwich, Steven D.
Journal Title
Journal ISSN
Volume Title
Publisher
AAAI
Published Version
Research Projects
Organizational Units
Journal Issue
Abstract
The field of constraint acquisition (CA) aims to remove the “modelling bottleneck” by learning constraints from examples. However, it gives rise to a “data collection bottleneck” as humans must prepare a suitable (labelled) dataset. A recently published paper described an unsupervised CA method called MineAcq that can learn standard CA benchmarks. In this paper we summarise the results, and apply MineAcq to a new, noisy, unlabelled dataset that was not designed for CA.
Description
Keywords
Constraint acquisition (CA) , Artificial Intelligence (AI) , Constraint satisfaction problem (CSP) , MineAcq
Citation
Prestwich, S. D. (2022) 'Constraint Acquisition and the Data Collection Bottleneck', AAAI-22: The Thirty-Sixth AAAI Conference on Artificial Intelligence, Vancouver, BC, Canada, 22 Feb - 1 Mar.
Link to publisher’s version
https://aaai.org/Conferences/AAAI-22/
URI
https://hdl.handle.net/10468/14043
Collections
Computer Science - Conference Items
Insight Centre for Data Analytics - Conference Items
Copyright