Structural Landmarking and Interaction Modelling: on Resolution Dilemmas in Graph Classification
This work addresses interpretability and performance issues in graph classification for machine learning researchers, presenting a novel method that is not incremental but offers a new approach to modeling interactions.
The paper tackles the problem of graph classification by addressing the 'resolution dilemmas' that arise from mixing all graph components into a single vector, which affects interpretability and performance. The proposed SLIM model improves interpretability and accuracy by explicitly modeling interactions between graph parts, offering new insights in graph representation learning.
Graph neural networks are promising architecture for learning and inference with graph-structured data. Yet difficulties in modelling the ``parts'' and their ``interactions'' still persist in terms of graph classification, where graph-level representations are usually obtained by squeezing the whole graph into a single vector through graph pooling. From complex systems point of view, mixing all the parts of a system together can affect both model interpretability and predictive performance, because properties of a complex system arise largely from the interaction among its components. We analyze the intrinsic difficulty in graph classification under the unified concept of ``resolution dilemmas'' with learning theoretic recovery guarantees, and propose ``SLIM'', an inductive neural network model for Structural Landmarking and Interaction Modelling. It turns out, that by solving the resolution dilemmas, and leveraging explicit interacting relation between component parts of a graph to explain its complexity, SLIM is more interpretable, accurate, and offers new insight in graph representation learning.