Skip to main content
Aller à la page d’accueil de la Commission européenne (s’ouvre dans une nouvelle fenêtre)
français fr
CORDIS - Résultats de la recherche de l’UE
CORDIS

Foundations of Implicitly Constrained Learning: A Functional Framework

Objectif

Machine learning has thrived by training expressive models on large datasets, but is now shifting toward integrating prior knowledge—like physical laws or structure—into the learning process.
Many such problems can be formulated as an implicitly constrained learning (ICL) task, where a prediction model is subject to a partially known constraint whose unknown component is inferred by aligning model predictions with data.
ICL problems arise across domains: enforcing physics in scientific models, planning from a learned world model, or accounting for hidden variables for causal effect estimation.
These settings pose major challenges, especially when using expressive over-parameterized models like deep neural networks:
(1) the resulting optimization can become ill-defined;
(2) current methods (e.g. differential programming, penalization approaches, neural operators) are often unstable, inaccurate, or inefficient;
(3) learning theory for ICL is largely under-developed, limiting understanding of generalization and robustness.

The ambition of DELPHI is to establish a principled and efficient methodological framework for solving ICL,
capable of handling complex implicit constraints—ranging from auxiliary optimality conditions to partial differential equations and dynamical systems.
Our approach adopts a functional viewpoint that reframes ICL as a well-defined yet abstract problem, using deep networks only as approximation tools.

The functional viewpoint on ICL will drive theoretical, methodological and practical advances in machine learning, and unlock transformative advances in scientific discovery, causal reasoning, and intelligent decision-making. By reconciling the use of expressive neural approximations with implicit constraints, DELPHI will pave the way toward more reliable, efficient, and principled machine learning.

Champ scientifique (EuroSciVoc)

Ce projet n'a pas encore été classé par EuroSciVoc.
Proposez les domaines scientifiques qui vous semblent les plus pertinents et aidez-nous à améliorer notre service de classification.

Programme(s)

Thème(s)

Régime de financement

HORIZON-ERC - HORIZON ERC Grants

Voir tous les projets financés dans le cadre de ce programme de financement

Appel à propositions

(s’ouvre dans une nouvelle fenêtre) ERC-2026-STG

Voir tous les projets financés au titre de cet appel

Institution d’accueil

INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET AUTOMATIQUE
Contribution nette de l'UE
€ 1 495 291,00
Adresse
DOMAINE DE VOLUCEAU ROCQUENCOURT
78153 Le Chesnay Cedex
France
Région
Ile-de-France Ile-de-France Yvelines
Type d’activité
Research Organisations
Liens
Coût total
€ 1 497 291,25

Bénéficiaires (1)