Integrating multiple information sources and evaluating joint impact in network inference

Develop methods to effectively incorporate diverse node attributes and multiple interaction types within attributed multilayer networks and rigorously evaluate their collective impact on downstream network inference tasks.

Background

Existing probabilistic models for attributed networks typically handle only single-layer or homogeneous data types (e.g., one interaction type and one categorical attribute), making it difficult to represent real-world scenarios with heterogeneous information. The authors highlight that combining various sources of information and assessing their joint effect on inference tasks (such as community detection and prediction) remains unresolved in the literature, motivating their development of a more flexible approach.

References

As a consequence, addressing the challenge of effectively incorporating various sources of information and evaluating their collective impact on downstream network inference tasks remains an open issue.

— Flexible inference in heterogeneous and attributed multilayer networks  (2405.20918 - Contisciani et al., 2024) in Introduction

Besides the generalization along the centrality dimension, several open questions of the framework deserve investigation: (1) adaptive learning mechanisms for the fusion weights \omega{C}(\cdot,\cdot)---for example, end-to-end optimization of the weighting-function parameters under supervision from network topology or node attributes, upgrading empirical values'' toautomatic tuning''; (2) the generalization of the fusion paradigm to directed, weighted, and dynamic temporal networks; (3) scalability verification and approximation-algorithm design on large-scale complex networks ($|V|>104$); (4) quantitative analysis of the normalization scheme for local similarity scores and the theoretical optimality of the weighting coefficients.

— Link prediction in complex networks via fusing node centrality and local similarity indices  (2609.09658 - Zhang et al., 9 Sep 2026) in Section 7.2, Methodological outlook