Capturing the full complexity of temporal interactions

Determine modelling frameworks in network science that successfully capture the full complexity of time-varying interactions in complex systems represented as temporal networks.

Background

The paper introduces Hyper Egocentric Temporal Neighbourhoods (HETN) and their signatures (HETNS), extending egocentric temporal analysis from simple graphs to hypergraphs so as to include second-order (triadic) interactions. Using these tools, the authors analyze ten datasets of social interactions, showing that second-order structures are crucial for distinguishing behaviors across datasets, individuals, and time.

Despite these advances, the authors emphasize a broader unresolved challenge in network science: modelling frameworks that can fully capture and reproduce the intrinsic complexity of temporal interactions remain elusive. This underscores the need for principled approaches that go beyond pairwise links and adequately represent the dynamics and heterogeneity observed in empirical temporal data.

References

Nevertheless, capturing and modelling all the complexity of temporal interactions is an open question in network science.

— Patterns in temporal networks with higher-order egocentric structures  (2402.03866 - Arregui-García et al., 2024) in Conclusions

Therefore, it is an essential open challenge to establish characterization methods for temporal evolution and benchmarks for dynamic collaborative graph learning~\citep{kim2025subgraph}.

— From Euclidean to Graph-Structured Data: A Survey of Collaborative Learning  (2609.02984 - Bourgerie et al., 2 Sep 2026) in Section 5.4, paragraph “Open challenges of data-based techniques”