Work

Neurosymbolic Constraint Learning: A Short Survey

Year: 2026

Type: conference-paper

Abstract: Constraint learning aims to induce the constraints governing a combinatorial problem from symbolic examples, rather than requiring them to be manually specified.Recent neurosymbolic approaches extend the traditional purely symbolic setting to examples that contain subsymbolic data, such as images, combining neural training with symbolic constraint learning in an end-to-end fashion.This survey introduces neurosymbolic constraint learning as a generalization of the purely symbolic setting and positions several neurosymbolic approaches within this setting.By comparing these approaches along five key dimensions, we identify commonalities and methodological trends in the literature.

Source: Lirias

Authors: Dewulf, Quinten, Seibt, Annegret, Odense, Simon, Zombori, Zsolt, De Raedt, Luc

Institution: KU Leuven

Language: English

Topic: Constraint Satisfaction and Optimization

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funders: European Commission, Fonds Wetenschappelijk Onderzoek, Vlaamse regering

Award: Neurosymbolic AI for Constraint Learning

OpenAlex ID: W7220166498

FWCI:

Cites:

Cited by:

Related to: 10