Doorgaan naar hoofdnavigatie Doorgaan naar zoeken Ga verder naar hoofdinhoud

Probabilistic Circuits with Constraints via Convex Optimization

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

Samenvatting

This work addresses integrating probabilistic propositional logic constraints into the distribution encoded by a probabilistic circuit (PC). PCs are a class of tractable models that allow efficient computations (such as conditional and marginal probabilities) while achieving state-of-the-art performance in some domains. The proposed approach takes both a PC and constraints as inputs, and outputs a new PC that satisfies the constraints. This is done efficiently via convex optimization without the need to retrain the entire model. Empirical evaluations indicate that the combination of constraints and PCs can have multiple use cases, including the improvement of model performance under scarce or incomplete data, as well as the enforcement of machine learning fairness measures into the model without compromising model fitness. We believe that these ideas will open possibilities for multiple other applications involving the combination of logics and deep probabilistic models.
Originele taal-2Engels
TitelMachine Learning and Knowledge Discovery in Databases. Research Track
SubtitelEuropean Conference, ECML PKDD 2024, Vilnius, Lithuania, September 9–13, 2024, Proceedings
RedacteurenAlbert Bifet, Jesse Davis, Tomas Krilavičius, Meelis Kull, Eirini Ntoutsi, Indré Žliobaitė
UitgeverijSpringer
Pagina's161-177
Volume2
ISBN van elektronische versie978-3-031-70352-2
ISBN van geprinte versie978-3-031-70351-5
DOI's
StatusGepubliceerd - 2024
EvenementEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024 - Vilnius, Litouwen
Duur: 9 sep. 202413 sep. 2024

Publicatie series

NaamLecture Notes in Computer Science
Volume14943
NaamLecture Notes in Artificial Intellingence
Volume14943

Congres

CongresEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024
Land/RegioLitouwen
StadVilnius
Periode9/09/2413/09/24

Vingerafdruk

Duik in de onderzoeksthema's van 'Probabilistic Circuits with Constraints via Convex Optimization'. Samen vormen ze een unieke vingerafdruk.

Citeer dit