On Discprecncies between Perturbation Evaluations of Graph Neural Network Attributions

• URL: http://arxiv.org/abs/2401.00633v1
• Date: Mon, 1 Jan 2024 02:03:35 GMT
• Title: On Discprecncies between Perturbation Evaluations of Graph Neural Network Attributions
• Authors: Razieh Rezaei, Alireza Dizaji, Ashkan Khakzar, Anees Kazi, Nassir Navab, Daniel Rueckert
• Abstract summary: We assess attribution methods from a perspective not previously explored in the graph domain: retraining. The core idea is to retrain the network on important (or not important) relationships as identified by the attributions. We run our analysis on four state-of-the-art GNN attribution methods and five synthetic and real-world graph classification datasets.
• Score: 49.8110352174327
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Neural networks are increasingly finding their way into the realm of graphs and modeling relationships between features. Concurrently graph neural network explanation approaches are being invented to uncover relationships between the nodes of the graphs. However, there is a disparity between the existing attribution methods, and it is unclear which attribution to trust. Therefore research has introduced evaluation experiments that assess them from different perspectives. In this work, we assess attribution methods from a perspective not previously explored in the graph domain: retraining. The core idea is to retrain the network on important (or not important) relationships as identified by the attributions and evaluate how networks can generalize based on these relationships. We reformulate the retraining framework to sidestep issues lurking in the previous formulation and propose guidelines for correct analysis. We run our analysis on four state-of-the-art GNN attribution methods and five synthetic and real-world graph classification datasets. The analysis reveals that attributions perform variably depending on the dataset and the network. Most importantly, we observe that the famous GNNExplainer performs similarly to an arbitrary designation of edge importance. The study concludes that the retraining evaluation cannot be used as a generalized benchmark and recommends it as a toolset to evaluate attributions on a specifically addressed network, dataset, and sparsity.

Related papers

• Deep Generative Models for Subgraph Prediction [10.56335881963895]
  This paper introduces subgraph queries as a new task for deep graph learning. Subgraph queries jointly predict the components of a target subgraph based on evidence that is represented by an observed subgraph. We utilize a probabilistic deep Graph Generative Model to answer subgraph queries.
  arXiv  Detail & Related papers  (2024-08-07T19:24:02Z)
• GNN-LoFI: a Novel Graph Neural Network through Localized Feature-based Histogram Intersection [51.608147732998994]
  Graph neural networks are increasingly becoming the framework of choice for graph-based machine learning. We propose a new graph neural network architecture that substitutes classical message passing with an analysis of the local distribution of node features.
  arXiv  Detail & Related papers  (2024-01-17T13:04:23Z)
• State of the Art and Potentialities of Graph-level Learning [54.68482109186052]
  Graph-level learning has been applied to many tasks including comparison, regression, classification, and more. Traditional approaches to learning a set of graphs rely on hand-crafted features, such as substructures. Deep learning has helped graph-level learning adapt to the growing scale of graphs by extracting features automatically and encoding graphs into low-dimensional representations.
  arXiv  Detail & Related papers  (2023-01-14T09:15:49Z)
• Classification of vertices on social networks by multiple approaches [1.370151489527964]
  In the case of social networks, it is crucial to evaluate the labels of discrete communities. For each of these interaction-based entities, a social graph, a mailing dataset, and two citation sets are selected as the testbench repositories. This paper was not only assessed the most valuable method but also determined how graph neural networks work.
  arXiv  Detail & Related papers  (2023-01-13T09:42:55Z)
• Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications [52.65389473899139]
  Graphs are a powerful tool for representing and analyzing unstructured, non-Euclidean data ubiquitous in the healthcare domain. Recent works have shown that considering relationships between input data samples have a positive regularizing effect for the downstream task. We propose Graph-in-Graph (GiG), a neural network architecture for protein classification and brain imaging applications.
  arXiv  Detail & Related papers  (2022-04-01T10:01:37Z)
• Temporal Graph Network Embedding with Causal Anonymous Walks Representations [54.05212871508062]
  We propose a novel approach for dynamic network representation learning based on Temporal Graph Network. For evaluation, we provide a benchmark pipeline for the evaluation of temporal network embeddings. We show the applicability and superior performance of our model in the real-world downstream graph machine learning task provided by one of the top European banks.
  arXiv  Detail & Related papers  (2021-08-19T15:39:52Z)
• Graph Belief Propagation Networks [34.137798598227874]
  We introduce a model that combines the advantages of graph neural networks and collective classification. In our model, potentials on each node only depend on that node's features, and edge potentials are learned via a coupling matrix. Our approach can be viewed as either an interpretable message-passing graph neural network or a collective classification method with higher capacity and modernized training.
  arXiv  Detail & Related papers  (2021-06-06T05:24:06Z)
• Towards Deeper Graph Neural Networks [63.46470695525957]
  Graph convolutions perform neighborhood aggregation and represent one of the most important graph operations. Several recent studies attribute this performance deterioration to the over-smoothing issue. We propose Deep Adaptive Graph Neural Network (DAGNN) to adaptively incorporate information from large receptive fields.
  arXiv  Detail & Related papers  (2020-07-18T01:11:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.