1.8k citations
- Centre National de la Recherche ScientifiqueFR449 papers
- Laboratoire d'Informatique Gaspard-MongeFR267 papers
- Laboratoire d’Analyse et de Mathématiques AppliquéesFR175 papers
- Université Paris-Est CréteilFR174 papers
- Université Paris CitéFR95 papers
- Institut des Sciences de la TerreFR91 papers
- Université Grenoble AlpesFR88 papers
- Lyon 1 UniversitéFR66 papers
- Université Savoie Mont BlancFR66 papers
- Sorbonne UniversitéFR61 papers
- Université Paris-SaclayFR48 papers
- Université Paris Sciences et LettresFR44 papers
Showing 2025 · cs.LGShow all
3 papers · 2 filters
cs.LG2025
Network-Wide Traffic Volume Estimation from Speed Profiles using a Spatio-Temporal Graph Neural Network with Directed Spatial Attention
Léo Hein, Giovanni de Nunzio, Giovanni Chierchia +2
Existing traffic volume estimation methods typically address either forecasting traffic on sensor-equipped roads or spatially imputing missing volumes using nearby sensors. While f…
cs.LG2025
Sliced ReLU attention: Quasi-linear contextual expressivity via sorting
François-Xavier Vialard, Siwan Boufadène
We introduce sliced ReLU attention, a new attention mechanism that departs structurally from both softmax and its approximation alternatives. Instead of applying a nonlinearity to…
cs.LG2025
The late-stage training dynamics of (stochastic) subgradient descent on homogeneous neural networks
Sholom Schechtman, Nicolas Schreuder
We analyze the implicit bias of constant step stochastic subgradient descent (SGD). We consider the setting of binary classification with homogeneous neural networks - a large clas…