Distinguished In Uniform: Self Attention Vs. Virtual Nodes

1 code implementation • 20 May 2024 • Eran Rosenbluth, Jan Tönshoff, Martin Ritzert, Berke Kisin, Martin Grohe

We first clarify that this form of universality is not unique to GTs: Using the same positional encodings, also pure MPGNNs and even 2-layer MLPs are non-uniform universal approximators.

One Model, Any CSP: Graph Neural Networks as Fast Global Search Heuristics for Constraint Satisfaction

1 code implementation • 22 Aug 2022 • Jan Tönshoff, Berke Kisin, Jakob Lindner, Martin Grohe

We propose a universal Graph Neural Network architecture which can be trained as an end-2-end search heuristic for any Constraint Satisfaction Problem (CSP).

Combinatorial Optimization Graph Neural Network