Where mathematics becomes executable innovation.
Explore the open repository for the project algorithms, software environment, and reproducible computational research and tools connecting at the interface of Control and Machine Learning. From robust deep-learning methods and federated learning to hybrid ResNet/PDE models. Designed for reproducible research and real-world impact, CoDeFeL equips researchers and practitioners to address challenges in digital health, autonomous driving, energy networks, recommendation systems, and beyond.
Working Package 1
Asymptotics and Turnpike for Deep Neural Networks
Clustering in pure-attention hardmax transformers
Clustering dynamics in pure-attention hardmax Transformers and their role in classification tasks, with a particular focus on sentiment analysis.
Author: Albert Alcalde & Enrique Zuazua
Language: Python
Nov. 6, 2024
Working Package 2
Complexity of ResNet Dynamics
Attention rollout for Machine Learning interpretability
Attention-mediated propagation in tabular Transformers through attention rollout.
Author: Umberto Biccari & Enrique Zuazua
Language: Python
Jul. 24, 2026
SHAP values for multi-output ML models
SHAP values-based interpretability of multi-output Machine Learning models.
Author: Umberto Biccari, Roberto Morales & Enrique Zuazua
Language: Python
Feb. 26, 2026
Fourier SHAP values
Python implementation of Fourier-based SHAP values for explaining neural network predictions in a biomedical classification task.
Author: Roberto Morales
Language: Python
Oct. 31, 2025
Spiking Neural Networks
Training of a basic Spiking Neural Network architecture.
Author: Umberto Biccari
Language: Python
Sep. 26, 2025
Working Package 3
Federated Learning
A potential game perspective in Federated Learning
Implementation of a potential game strategy for Federated Learning.
Author: Kang Liu, Ziqi Wang & Enrique Zuazua
Language: Python
Jun. 1, 2026
Decentralized Learning with coordination constraints
A source code to implement Decentralized Learning as a Multi-Objective Optimization problem.
Author: Roberto Morales & Umberto Biccari
Language: Python
Jul. 18, 2025
FedADMM-InSa
Inexact and self-adaptive Alternating Direction Method of Multipliers (ADMM) for Federated Learning.
Author: Yongcun Song, Ziqi Wang & Enrique Zuazua
Language: Python
Oct. 1, 2024
Working Package 4
Modelling through Control and Machine Learning
bi-HYCO: Cooperative PDE learning with fragmented observations
Implementation of the Bi-HYCO framework for PDE parameter identification with fragmented observations.
Author: Umberto Biccari, Jun Chen, Roberto Morales & Enrique Zuazua
Language: Python
Sep. 4, 2026
Riccati solution via DeepONet
Training of Deep Operator Networks for the resolution of Algebraic and Differential Riccati equations.
Author: Jun Chen & Umberto Biccari
Language: Python
Apr. 21, 2026
DeepONet for the stabilization of reaction-diffusion systems
Training of Deep Operator Networks for the stabilization of reaction-diffusion systems.
Author: Kaijing Lyu & Umberto Biccari
Language: Python
Feb. 26, 2026
DeepONet for the stabilization of stochastic PDE-ODE systems
Training of Deep Operator Networks for the stabilization of stochastic PDE-ODE systems.
Author: Kaijing Lyu & Umberto Biccari
Language: Python
Ago. 5, 2025
Semi-autonomous Neural ODEs
A source code to approximate the behavior of dynamical systems using neural networks.
Author: Ziqian Li, Kang Liu, Lorenzo Liverani & Enrique Zuazua
Language: Python
Jan. 11, 2025
HYCO: The hybrid-collapse strategy for PDEs
Implementation of the HYCO strategy for the hybrid modelling of PDEs.
Author: Lorenzo Liverani, Matthys Steynberg & Enrique Zuazua
Language: Python
Dec. 21, 2024