Luca Franceschi

ReFactorGNNs: Revisiting Factorisation-based Models from a Message-Passing Perspective

Jul 21, 2022
Yihong Chen, Pushkar Mishra, Luca Franceschi, Pasquale Minervini, Pontus Stenetorp, Sebastian Riedel

Figure 2 for ReFactorGNNs: Revisiting Factorisation-based Models from a Message-Passing Perspective

Figure 3 for ReFactorGNNs: Revisiting Factorisation-based Models from a Message-Passing Perspective

Figure 4 for ReFactorGNNs: Revisiting Factorisation-based Models from a Message-Passing Perspective

Access Paper or Ask Questions

Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

Jun 03, 2021
Mathias Niepert, Pasquale Minervini, Luca Franceschi

Figure 2 for Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

Figure 3 for Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

Figure 4 for Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

Access Paper or Ask Questions

On the Iteration Complexity of Hypergradient Computation

Jul 10, 2020
Riccardo Grazzi, Luca Franceschi, Massimiliano Pontil, Saverio Salzo

Figure 2 for On the Iteration Complexity of Hypergradient Computation

Figure 3 for On the Iteration Complexity of Hypergradient Computation

Figure 4 for On the Iteration Complexity of Hypergradient Computation

* accepted at ICML 2020; 19 pages, 4 figures; code at https://github.com/prolearner/hypertorch (corrected typos and one reference)

Access Paper or Ask Questions

Scheduling the Learning Rate via Hypergradients: New Insights and a New Algorithm

Oct 18, 2019
Michele Donini, Luca Franceschi, Massimiliano Pontil, Orchid Majumder, Paolo Frasconi

Figure 2 for Scheduling the Learning Rate via Hypergradients: New Insights and a New Algorithm

Figure 3 for Scheduling the Learning Rate via Hypergradients: New Insights and a New Algorithm

Figure 4 for Scheduling the Learning Rate via Hypergradients: New Insights and a New Algorithm

* Code available at https://github.com/awslabs/adatune

Access Paper or Ask Questions

Learning Discrete Structures for Graph Neural Networks

May 17, 2019
Luca Franceschi, Mathias Niepert, Massimiliano Pontil, Xiao He

Figure 2 for Learning Discrete Structures for Graph Neural Networks

Figure 3 for Learning Discrete Structures for Graph Neural Networks

Figure 4 for Learning Discrete Structures for Graph Neural Networks

* To appear as a conference paper at ICML 2019, code at https://github.com/lucfra/LDS

Access Paper or Ask Questions

Fast and Continuous Foothold Adaptation for Dynamic Locomotion through CNNs

Feb 15, 2019
Octavio Villarreal, Victor Barasuol, Marco Camurri, Luca Franceschi, Michele Focchi, Massimiliano Pontil, Darwin G. Caldwell, Claudio Semini

Figure 2 for Fast and Continuous Foothold Adaptation for Dynamic Locomotion through CNNs

Figure 3 for Fast and Continuous Foothold Adaptation for Dynamic Locomotion through CNNs

Figure 4 for Fast and Continuous Foothold Adaptation for Dynamic Locomotion through CNNs

* 9 pages, 11 figures. Accepted to RA-L + ICRA 2019, January 2019

Access Paper or Ask Questions

Bilevel Programming for Hyperparameter Optimization and Meta-Learning

Jul 03, 2018
Luca Franceschi, Paolo Frasconi, Saverio Salzo, Riccardo Grazzi, Massimilano Pontil

Figure 2 for Bilevel Programming for Hyperparameter Optimization and Meta-Learning

Figure 3 for Bilevel Programming for Hyperparameter Optimization and Meta-Learning

Figure 4 for Bilevel Programming for Hyperparameter Optimization and Meta-Learning

* ICML 2018; code for replicating experiments at https://github.com/prolearner/hyper-representation, main package (Far-HO) at https://github.com/lucfra/FAR-HO

Access Paper or Ask Questions

Far-HO: A Bilevel Programming Package for Hyperparameter Optimization and Meta-Learning

Jun 13, 2018
Luca Franceschi, Riccardo Grazzi, Massimiliano Pontil, Saverio Salzo, Paolo Frasconi

Figure 2 for Far-HO: A Bilevel Programming Package for Hyperparameter Optimization and Meta-Learning

* This submission is a reduced version of (Franceschi et al., arXiv:1806.04910) which has been accepted at the main ICML 2018 conference. In this paper we illustrate the software framework, material that could not be included in the conference paper

Access Paper or Ask Questions

1

2

>>