2021年10月26日 · Abstract: The interdependence between nodes in graphs is key to improve class predictions on nodes and utilized in approaches like Label Propagation (LP) or in Graph Neural Networks (GNN). Nonetheless, uncertainty estimation for non-independent node-level predictions is under-explored.

Our main Graph Posterior Network model can be found in gpn.models.gpn_base.py. Ablated models can be found in a similar fashion, i.e. PostNet in gpn.models.gpn_postnet.py, PostNet+diffusion in gpn.models.gpn_postnet_diff.py and the model diffusiong log-beta scores in gpn.models.gpn_log_beta.py.

In this work, we explore uncertainty quantification for node classification in three ways: (1) We derive three axioms explicitly characterizing the expected predictive uncertainty behavior in homophilic attributed graphs.(2) We propose a new model Graph Posterior Network (GPN) which explicitly performs Bayesian posterior updates for predictions ...

The interdependence between nodes in graphs is key to improve class predictions on nodes and utilized in approaches like Label Propagation (LP) or in Graph Neural Networks (GNNs). Nonetheless, uncertainty estimation for non-independent node-level predictions is under-explored.

In this work, we explore uncertainty quantification for node classification in three ways: (1) We derive three axioms explicitly characterizing the expected predictive uncertainty behavior in homophilic attributed graphs.(2) We propose a new model Graph Posterior Network (GPN) which explicitly performs Bayesian posterior updates for predictions ...

The interdependence between nodes in graphs is key to improve class predictions on nodes and utilized in approaches like Label Propagation (LP) or in Graph Neural Networks (GNN). Nonetheless, uncertainty estimation for non-independent node-level predictions is under-explored.

2021年9月28日 · The interdependence between nodes in graphs is key to improve class predictions on nodes and utilized in approaches like abel Propagation (LP) or in Graph Neural Networks (GNNs). Nonetheless, uncertainty estimation for non-independent node-level predictions is under-explored.

The interdependence between nodes in graphs is key to improve class predictions on nodes and utilized in approaches like abel Propagation (LP) or in Graph Neural Networks (GNNs). Nonetheless, uncertainty estimation for non-independent node-level predictions is under-explored.

2021年10月26日 · A new model Graph Posterior Network (GPN) is proposed which explicitly performs Bayesian posterior updates for predictions on interdependent nodes and outperforms existing approaches for uncertainty estimation in the experiments.

In this work, we make the first attempt to endow a well-trained GNN with the OOD detection ability without modifying its param-eters. To this end, we design a post-hoc framework with Adaptive Amplifier for Graph OOD Detection, named AAGOD, concentrating on data-centric manipulation.