Poster
in
Workshop: Differentiable Almost Everything: Differentiable Relaxations, Algorithms, Operators, and Simulators

A framework for differentiable Supervised Graph Prediction

Paul KRZAKALA ⋅ Junjie Yang ⋅ Rémi Flamary ⋅ Florence d'Alché-Buc ⋅ Charlotte Laclau ⋅ Matthieu Labeau

Keywords: Deep Learning Optimal Transport Differentiable Graph Prediction Graphs

Project Page [ OpenReview]

Abstract

We introduce a general framework which enable to train a deep neural network to predict graph. The framework is built upon a novel Optimal Transport loss that exhibits all necessary properties (permutation invariance and differentiability) and is designed to handle graphs of any size. We showcase the versatility and state-of-the-art performances of the proposed approach on a variety of real-world tasks and a novel challenging synthetic dataset.

Chat is not available.