🚨 Acceptance Decisions Are In! 🚨
We’ve finalized the acceptance decisions for the submissions received before the priority deadline. Make sure to check your inbox!

✨ We’re still accepting submissions on a rolling basis! Submit your abstract here: tinyurl.com/winsNetSci25

Check out the code on our repository! (5/5)

github.com/lenafm/metab...

We showcase how metablox works on a number of synthetic and empirical networks, and that it can be used in comparative settings: if you have an entire collection of networks with shared metadata OR a network with multiple sets of categorical metadata. (4/5)

In our paper, we introduce the metablox tool, to quantify the strength and type of relationship between categorical node metadata and the block structure of a network (using Stochastic block models and description length). (3/5)

Different sets of node metadata may (a) relate to a network’s block structure to varying degrees, and (b) resemble different types of arrangements, such as community structure or core-periphery structure! (2/5)

Our model serves as a benchmark for future research, to gauge how mesoscale structure detection algorithms deal with mesoscale ambiguity + we emphasise the importance of considering coexisting structures. (6/6)

In our experimental set-up, the coexistence of the two partitions was only detected in a small number of cases — mostly one (dominating) structure was preferred! (5/6)

We showcase the SCBM by generating networks with coexisting bi-community and core-periphery structure (with varying ‘structural strength’) and we explore how well they are picked up by different SBM variants. (4/6)

We introduce the Stochastic Cross-Block Model (SCBM), a generative framework for networks with multiple coexisting ground-truth partitions. (3/6)

Real networks often feature multiple coexisting structures which community detection algorithms may or may not pick up, calling for a benchmark network model with built-in ambiguity on the mesoscale. (2/6)

Our model serves as a benchmark for future research, to gauge how mesoscale structure detection algorithms deal with mesoscale ambiguity + we emphasise the importance of considering coexisting structures. (6/6)

In our experimental set-up, the coexistence of the two partitions was only detected in a small number of cases — mostly one (dominating) structure was preferred! (5/6)

We showcase the SCBM by generating networks with coexisting bi-community and core-periphery structure (with varying ‘structural strength’) and we explore how well they are picked up by different SBM variants. (4/6)