SubsidieMeestersControl for Deep and Federated Learning
CoDeFeL aims to enhance machine learning methods through control theory, developing efficient ResNet architectures and federated learning techniques for applications in digital medicine and recommendations.
Projectdetails
Introduction
Machine Learning (ML) is forging a new era in Applied Mathematics (AM), leading to innovative and powerful methods. But the need for theoretical guarantees generates challenging, fundamental, deep mathematical questions.
Addressing the Challenge
This great challenge can be addressed from the perspective of other, more mature areas of AM. CoDeFeL seeks to do so from the rich interface between Control Theory (CT) and ML, contributing to the analytical foundations of ML methods, significantly enlarging, and updating the range of applications of CT.
Recent Results
As our recent results show, classification, regression, and prediction problems in Supervised Learning (SL) and the Universal Approximation Theorem can be successfully recast as the simultaneous or ensemble controllability property of Residual Neural Networks (ResNets).
Development of ResNets
Following this path, we will develop ResNets of minimal complexity and cost, addressing the deep, intricate issue of linking the structure of the data set to be classified with the dynamics of the networks trained.
New Architectures
Taking the turnpike principle as our inspiration, we will build new simplified ResNet architectures. This, however, raises major challenges for the genuinely nonlinear dynamics that ResNets represent.
Adjoint Methods
Adjoint methods will also be developed and applied to understand the sensitivity of ResNets, proposing techniques for Adversarial Training and computing Saliency Maps, applicable in Unsupervised Learning.
Application Areas
The project is strongly inspired by the challenges arising in relevant applications in digital medicine and internet recommendation systems, among other areas.
Hybrid Methods
Accordingly, we will also develop a body of rich, hybrid, cutting-edge methods for data-aware modelling, combining ResNet surrogate models and those inspired by Mechanics, with the aid of Model Predictive Control strategies.
Federated Learning
New Federated Learning methodologies with privacy preservation guarantees will also be developed.
Repository
The computational counterparts will be brought together in a new CoDeFeL GitHub repository.
Financiële details & Tijdlijn
Financiële details
| Subsidiebedrag | € 2.499.224 |
| Totale projectbegroting | € 2.499.224 |
Tijdlijn
| Startdatum | 1-9-2024 |
| Einddatum | 31-8-2029 |
| Subsidiejaar | 2024 |
Partners & Locaties
Projectpartners
- FRIEDRICH-ALEXANDER-UNIVERSITAET ERLANGEN-NUERNBERGpenvoerder
- UNIVERSIDAD DE LA IGLESIA DE DEUSTO ENTIDAD RELIGIOSA
Land(en)
Vergelijkbare projecten binnen European Research Council
| Project | Regeling | Bedrag | Jaar | Actie |
|---|---|---|---|---|
Reconciling Classical and Modern (Deep) Machine Learning for Real-World ApplicationsAPHELEIA aims to create robust, interpretable, and efficient machine learning models that require less data by integrating classical methods with modern deep learning, fostering interdisciplinary collaboration. | ERC Consolid... | € 1.999.375 | 2023 | Details |
A Theory of Neural Networks for ControlDevelop a comprehensive mathematical theory of neural networks for control to enhance safety, robustness, and reliability in critical applications like healthcare and aerospace. | ERC Starting... | € 1.493.750 | 2024 | Details |
Dynamics-Aware Theory of Deep LearningThis project aims to create a robust theoretical framework for deep learning, enhancing understanding and practical tools to improve model performance and reduce complexity in various applications. | ERC Starting... | € 1.498.410 | 2022 | Details |
Continual and Sequential Learning for Artificial IntelligenceDevelop a Continual and Sequential Learning AI that integrates RL with advanced data gathering to autonomously adapt and explore in dynamic environments for scientific breakthroughs. | ERC Starting... | € 1.259.375 | 2025 | Details |
Using deep learning to understand computations in neural circuits with Connectome-constrained Mechanistic ModelsThis project aims to develop a machine learning framework that integrates mechanistic modeling and deep learning to understand neural computations in Drosophila melanogaster's circuits. | ERC Consolid... | € 1.997.321 | 2023 | Details |
Reconciling Classical and Modern (Deep) Machine Learning for Real-World Applications
APHELEIA aims to create robust, interpretable, and efficient machine learning models that require less data by integrating classical methods with modern deep learning, fostering interdisciplinary collaboration.
A Theory of Neural Networks for Control
Develop a comprehensive mathematical theory of neural networks for control to enhance safety, robustness, and reliability in critical applications like healthcare and aerospace.
Dynamics-Aware Theory of Deep Learning
This project aims to create a robust theoretical framework for deep learning, enhancing understanding and practical tools to improve model performance and reduce complexity in various applications.
Continual and Sequential Learning for Artificial Intelligence
Develop a Continual and Sequential Learning AI that integrates RL with advanced data gathering to autonomously adapt and explore in dynamic environments for scientific breakthroughs.
Using deep learning to understand computations in neural circuits with Connectome-constrained Mechanistic Models
This project aims to develop a machine learning framework that integrates mechanistic modeling and deep learning to understand neural computations in Drosophila melanogaster's circuits.