activity
20242026
collaborators
Showing cs.LGShow all

5 papers · 1 filter

cs.LG2026

N-vium: Mixture-of-Exits Transformer for Accelerated Exact Generation

Aleksander Lorenc, Frédéric Berdoz, Joël Mathys +1

Improving the inference efficiency of autoregressive transformers typically means reducing FLOPs per token, usually through approximations that degrade model quality. We introduce…

cs.LG2026

From Message-Passing to Linearized Graph Sequence Models

Joël Mathys, Basil Rohner, Saku Peltonen +1

Message-passing based approaches form the default backbone of most learning architectures on graph-structured data. However, the rapid progress of modern deep learning architecture…

cs.LG2025

Learn to Jump: Adaptive Random Walks for Long-Range Propagation through Graph Hierarchies

Joël Mathys, Federico Errica

Message-passing architectures struggle to sufficiently model long-range dependencies in node and graph prediction tasks. We propose a novel approach exploiting hierarchical graph s…

cs.LG2025

Beyond Interpolation: Extrapolative Reasoning with Reinforcement Learning and Graph Neural Networks

Niccolò Grillo, Andrea Toccaceli, Joël Mathys +3

Despite incredible progress, many neural architectures fail to properly generalize beyond their training distribution. As such, learning to reason in a correct and generalizable wa…

cs.LG2024

Flood and Echo Net: Algorithmically Aligned GNNs that Generalize

Joël Mathys, Florian Grötschla, Kalyan Varma Nadimpalli +1

Most Graph Neural Networks follow the standard message-passing framework where, in each step, all nodes simultaneously communicate with each other. We want to challenge this paradi…