arXiv subjects

Martin Rosvall

Publications and source records attributed to Martin Rosvall.

At least 19 recordsLinked to original sources

arXiv2026

Constrained user-item allocation for e-commerce marketing campaigns

When running marketing campaigns, retailers must decide which products to promote and which users to target. These decisions are inherently coupled: effective campaigns match users and items with strong mutual affinity into non-overlapping groups of predefined sizes. However, existing approaches assume predefined campaign structure or decouple item selection from user assignment, and cannot discover campaign groupings directly from joint interaction patterns. We therefore formalize this campaign problem as auto-targeting: jointly selecting users and items to construct multiple disjoint campaigns. To solve this combinatorial problem, we propose three complementary strategies: (i) constrained spectral biclustering to find dense regions in the user-item affinity matrix, (ii) greedy local search with pairwise swaps for combinatorial refinement, and (iii) a multi-armed bandit framework to escape local optima through exploration. We evaluate these methods on a synthetic dataset, the Amazon Reviews benchmarks, and large-scale proprietary commercial data, and compare the results to simulated annealing as a baseline. The results show that biclustering consistently achieves the highest campaign quality, lift, and fairness scores. While biclustering runs efficiently on smaller datasets, its runtime increases substantially on very large ones, where bandit-based methods instead offer a scalable alternative.

cs.LG↗

arXiv2025

Reliable data clustering with Bayesian community detection

From neuroscience and genomics to systems biology and ecology, researchers rely on clustering similarity data to uncover modular structure. Yet widely used clustering methods, such as hierarchical clustering, k-means, and WGCNA, lack principled model selection, leaving them susceptible to noise. A common workaround sparsifies a correlation matrix representation to remove noise before clustering, but this extra step introduces arbitrary thresholds that can distort the structure and lead to unreliable results. To detect reliable clusters, we capitalize on recent advances in network science to unite sparsification and clustering with principled model selection. We test two Bayesian community detection methods, the Degree-Corrected Stochastic Block Model and the Regularized Map Equation, both grounded in the Minimum Description Length principle for model selection. In synthetic data, they outperform traditional approaches, detecting planted clusters under high-noise conditions and with fewer samples. Compared to WGCNA on gene co-expression data, the Regularized Map Equation identifies more robust and functionally coherent gene modules. Our results establish Bayesian community detection as a principled and noise-resistant framework for uncovering modular structure in high-dimensional data across fields.

stat.ML↗

arXiv2025

Deep Graph Learning will stall without Network Science

Deep graph learning focuses on flexible and generalizable models that learn patterns in an automated fashion. Network science focuses on models and measures revealing the organizational principles of complex systems with explicit assumptions. Both fields share the same goal: to better model and understand patterns in graph-structured data. However, deep graph learning prioritizes empirical performance but ignores fundamental insights from network science. Our position is that deep graph learning will stall without insights from network science. In this position paper, we formulate six Calls for Action to leverage untapped insights from network science to address current issues in deep graph learning, ensuring the field continues to make progress.

cs.LG↗

arXiv2025

Concise network models of memory dynamics reveal explainable patterns in path data

Networks are a powerful tool to model the structure and dynamics of complex systems across scales. Direct connections between system components are often represented as edges, while paths and walks capture indirect interactions. This approach assumes that flows in the system are sequences of independent transitions. Path data from real-world systems often have higher-order dependencies, which require more sophisticated models. In this work, we propose a method to construct concise networks from path data that interpolate between first and second-order models. We prioritise simplicity and interpretability by creating state nodes that capture latent modes of second-order effects and introducing an interpretable measure to balance model size and accuracy. In both synthetic and real-world applications, our method reveals large-scale memory patterns and constructs concise networks that provide insights beyond the first-order model at the fraction of the size of a second-order model.

physics.soc-ph↗

arXiv2024

Compressing regularized dynamics improves link prediction with the map equation in sparse networks

Predicting future interactions or novel links in networks is an indispensable tool across diverse domains, including genetic research, online social networks, and recommendation systems. Among the numerous techniques developed for link prediction, those leveraging the networks' community structure have proven highly effective. For example, the recently proposed MapSim predicts links based on a similarity measure derived from the code structure of the map equation, a community-detection objective function that operates on network flows. However, the standard map equation assumes complete observations and typically identifies many small modules in networks where the nodes connect through only a few links. This aspect can degrade MapSim's performance on sparse networks. To overcome this limitation, we propose to incorporate a global regularization method based on a Bayesian estimate of the transition rates along with three local regularization methods. The regularized versions of the map equation compensate for incomplete observations and mitigate spurious community fragmentation in sparse networks. The regularized methods outperform standard MapSim and several state-of-the-art embedding methods in highly sparse networks. This performance holds across multiple real-world networks with randomly removed links, simulating incomplete observations. Among the proposed regularization methods, the global approach provides the most reliable community detection and the highest link prediction performance across different network densities. The principled method requires no hyperparameter tuning and runs at least an order of magnitude faster than the embedding methods.

cs.SI↗

arXiv2024

A Unified Framework for Tabular Generative Modeling: Loss Functions, Benchmarks, and Improved Multi-objective Bayesian Optimization Approaches

Deep learning (DL) models require extensive data to achieve strong performance and generalization. Deep generative models (DGMs) offer a solution by synthesizing data. Yet current approaches for tabular data often fail to preserve feature correlations and distributions during training, struggle with multi-metric hyperparameter selection, and lack comprehensive evaluation protocols. We address this gap with a unified framework that integrates training, hyperparameter tuning, and evaluation. First, we introduce a novel correlation- and distribution-aware loss function that regularizes DGMs, enhancing their ability to generate synthetic tabular data that faithfully represents the underlying data distributions. Theoretical analysis establishes stability and consistency guarantees. To enable principled hyperparameter search via Bayesian optimization (BO), we also propose a new multi-objective aggregation strategy based on iterative objective refinement Bayesian optimization (IORBO), along with a comprehensive statistical testing framework. We validate the proposed approach using a benchmarking framework with twenty real-world datasets and ten established tabular DGM baselines. The correlation-aware loss function significantly improves synthetic data fidelity and downstream machine learning (ML) performance, while IORBO consistently outperforms standard Bayesian optimization (SBO) in hyperparameter selection. The unified framework advances tabular generative modeling beyond isolated method improvements. Code is available at: https://github.com/vuhoangminh/TabGen-Framework

cs.LG↗

arXiv2023

Community Detection with the Map Equation and Infomap: Theory and Applications

Real-world networks have a complex topology comprising many elements often structured into communities. Revealing these communities helps researchers uncover the organizational and functional structure of the system that the network represents. However, detecting community structures in complex networks requires selecting a community detection method among a multitude of alternatives with different network representations, community interpretations, and underlying mechanisms. This tutorial focuses on a popular community detection method called the map equation and its search algorithm Infomap. The map equation framework for community detection describes communities by analyzing dynamic processes on the network. Thanks to its flexibility, the map equation provides extensions that can incorporate various assumptions about network structure and dynamics. To help decide if the map equation is a suitable community detection method for a given complex system and problem at hand - and which variant to choose - we review the map equation's theoretical framework and guide users in applying the map equation to various research problems.

physics.soc-ph↗

arXiv2023

Infomap Bioregions 2 - Exploring the interplay between biogeography and evolution

Identifying and understanding the large-scale biodiversity patterns in time and space is vital for conservation and addressing fundamental ecological and evolutionary questions. Network-based methods have proven useful for simplifying and highlighting important structures in species distribution data. However, current network-based biogeography approaches cannot exploit the evolutionary information available in phylogenetic data. We introduce a method for incorporating evolutionary relationships into species occurrence networks to produce more biologically informative and robust bioregions. To keep the bipartite network structure where bioregions are grid cells indirectly connected through shared species, we incorporate the phylogenetic tree by connecting ancestral nodes to the grid cells where their descendant species occur. To incorporate the whole tree without destroying the spatial signal of narrowly distributed species or ancestral nodes, we weigh tree nodes by the geographic information they provide. For a more detailed analysis, we enable integration of the evolutionary relationships at a specific time in the tree. By sweeping through the phylogenetic tree in time, our method interpolates between finding bioregions based only on distributional data and finding spatially segregated clades, uncovering evolutionarily distinct bioregions at different time slices. We also introduce a way to segregate the connections between evolutionary branches at a selected time to enable exploration of overlapping evolutionarily distinct regions. We have implemented these methods in Infomap Bioregions, an interactive web application that makes it easy to explore the possibly hierarchical and fuzzy patterns of biodiversity on different scales in time and space.

q-bio.PE↗

arXiv2023

Compressing network populations with modal networks reveals structural diversity

Analyzing relational data consisting of multiple samples or layers involves critical challenges: How many networks are required to capture the variety of structures in the data? And what are the structures of these representative networks? We describe efficient nonparametric methods derived from the minimum description length principle to construct the network representations automatically. The methods input a population of networks or a multilayer network measured on a fixed set of nodes and output a small set of representative networks together with an assignment of each network sample or layer to one of the representative networks. We identify the representative networks and assign network samples to them with an efficient Monte Carlo scheme that minimizes our description length objective. For temporally ordered networks, we use a polynomial time dynamic programming approach that restricts the clusters of network layers to be temporally contiguous. These methods recover planted heterogeneity in synthetic network populations and identify essential structural heterogeneities in global trade and fossil record networks. Our methods are principled, scalable, parameter-free, and accommodate a wide range of data, providing a unified lens for exploratory analyses and preprocessing large sets of network samples.

physics.soc-ph↗

arXiv2023

Module-based regularization improves Gaussian graphical models when observing noisy data

Inferring relations from correlational data allows researchers across the sciences to uncover complex connections between variables for insights into the underlying mechanisms. The researchers often represent inferred relations using Gaussian graphical models, requiring regularization to sparsify the models. Acknowledging that the modular structure of the inferred network is often studied, we suggest module-based regularization to balance under- and overfitting. Compared with the graphical lasso, a standard approach using the Gaussian log-likelihood for estimating the regularization strength, this approach better recovers and infers modular structure in noisy synthetic and real data. The module-based regularization technique improves the usefulness of Gaussian graphical models in the many applications where they are employed.

physics.data-an↗

arXiv2023

Mapping memory-biased dynamics with compact models reveals overlapping communities in large networks

Many real-world systems, from social networks to protein-protein interactions and species distributions, exhibit overlapping flow-based communities that reflect their functional organisation. However, reliably identifying such overlapping flow-based communities requires higher-order relational data, which are often unavailable. To address this challenge, we capitalise on the flow model underpinning the representation-learning algorithm node2vec and model higher-order flows through memory-biased random walks on first-order networks. Instead of simulating these walks, we model their higher-order dynamic constraints with compact models and control model complexity with an information-theoretic approach. Using the map equation framework, we identify overlapping modules in the resulting higher-order networks. Our compact-model approach proves robust across synthetic benchmark networks, reveals interpretable overlapping communities in empirical networks, and scales to large networks.

physics.soc-ph↗

arXiv2023

Cross-validation of correlation networks using modular structure

Correlation networks derived from multivariate data appear in many applications across the sciences. These networks are usually dense and require sparsification to detect meaningful structure. However, current methods for sparsifying correlation networks struggle with balancing overfitting and underfitting. We propose a module-based cross-validation procedure to threshold these networks, making modular structure an integral part of the thresholding. We illustrate our approach using synthetic and real data and find that its ability to recover a planted partition has a step-like dependence on the number of data samples. The reward for sampling more varies non-linearly with the number of samples, with minimal gains after a critical point. A comparison with the well-established WGCNA method shows that our approach allows for revealing more modular structure in the data used here.

physics.soc-ph↗

arXiv2023

Mapping change in higher-order networks with multilevel and overlapping communities

New network models of complex systems use layers, state nodes, or hyperedges to capture higher-order interactions and dynamics. Simplifying how the higher-order networks change over time or depending on the network model would be easy with alluvial diagrams, which visualize community splits and merges between networks. However, alluvial diagrams were developed for networks with regular nodes assigned to non-overlapping flat communities. How should they be defined for nodes in layers, state nodes, or hyperedges? How can they depict multilevel, overlapping communities? Here we generalize alluvial diagrams to map change in higher-order networks and provide an interactive tool for anyone to generate alluvial diagrams. We use the alluvial generator to illustrate the effect of modeling network flows with memory in a citation network, distinguishing multidisciplinary from field-specific journals.

physics.soc-ph↗

arXiv2023

Similarity-based Link Prediction from Modular Compression of Network Flows

Node similarity scores are a foundation for machine learning in graphs for clustering, node classification, anomaly detection, and link prediction with applications in biological systems, information networks, and recommender systems. Recent works on link prediction use vector space embeddings to calculate node similarities in undirected networks with good performance. Still, they have several disadvantages: limited interpretability, need for hyperparameter tuning, manual model fitting through dimensionality reduction, and poor performance from symmetric similarities in directed link prediction. We propose MapSim, an information-theoretic measure to assess node similarities based on modular compression of network flows. Unlike vector space embeddings, MapSim represents nodes in a discrete, non-metric space of communities and yields asymmetric similarities in an unsupervised fashion. We compare MapSim on a link prediction task to popular embedding-based algorithms across 47 networks and find that MapSim's average performance across all networks is more than 7% higher than its closest competitor, outperforming all embedding methods in 11 of the 47 networks. Our method demonstrates the potential of compression-based approaches in graph representation learning, with promising applications in other graph learning tasks.

cs.SI↗

arXiv2022

Variable Markov dynamics as a multi-focal lens to map multi-scale complex networks

From traffic flows on road networks to electrical signals in brain networks, many real-world networks contain modular structures of different sizes and densities. In the networks where modular structures emerge due to coupling between nodes with similar dynamical functions, we can identify them using flow-based community detection methods. However, these methods implicitly assume that communities are dense or clique-like which can shatter sparse communities due to a field-of-view limit inherent in one-step dynamics. Taking multiple steps with shorter or longer Markov time enables us to effectively zoom in or out to capture small or long-range communities. However, zooming out to avoid the field-of-view limit comes at the expense of introducing or increasing a lower resolution limit. Here we relax the constant Markov time constraint and introduce variable Markov dynamics as a multi-focal lens to capture functional communities in networks with a higher range of scales. With variable Markov time, a random walker can keep one-step dynamics in dense areas to avoid the resolution limit and move faster in sparse areas to detect long-range modular structures and prevent the field-of-view limit. We analyze the performance of variable Markov time using the flow-based community detection method called the map equation. We have implemented the map equation with variable Markov time in the search algorithm Infomap without any complexity overhead and tested its performance on synthetic and real-world networks from different domains. Results show that it outperforms the standard map equation in networks with constrained structures and locally sparse regions. In addition, the method estimates the optimal Markov time and avoids parameter tuning.

physics.soc-ph↗

arXiv2022

Map Equation Centrality: Community-aware Centrality based on the Map Equation

To measure node importance, network scientists employ centrality scores that typically take a microscopic or macroscopic perspective, relying on node features or global network structure. However, traditional centrality measures such as degree centrality, betweenness centrality, or PageRank neglect the community structure found in real-world networks. To study node importance based on network flows from a mesoscopic perspective, we analytically derive a community-aware information-theoretic centrality score based on network flow and the coding principles behind the map equation: map equation centrality. Map equation centrality measures how much further we can compress the network's modular description by not coding for random walker transitions to the respective node, using an adapted coding scheme and determining node importance from a network flow-based point of view. The information-theoretic centrality measure can be determined from a node's local network context alone because changes to the coding scheme only affect other nodes in the same module. Map equation centrality is agnostic to the chosen network flow model and allows researchers to select the model that best reflects the dynamics of the process under study. Applied to synthetic networks, we highlight how our approach enables a more fine-grained differentiation between nodes than node-local or network-global measures. Predicting influential nodes for two different dynamical processes on real-world networks with traditional and other community-aware centrality measures, we find that activating nodes based on map equation centrality scores tends to create the largest cascades in a linear threshold model.

cs.SI↗

arXiv2021

Mapping flows on weighted and directed networks with incomplete observations

Detecting significant community structure in networks with incomplete observations is challenging because the evidence for specific solutions fades away with missing data. For example, recent research shows that flow-based community detection methods can highlight spurious communities in sparse undirected and unweighted networks with missing links. Current Bayesian approaches developed to overcome this problem do not work for incomplete observations in weighted and directed networks that describe network flows. To address this gap, we extend the idea behind the Bayesian estimate of the map equation for unweighted and undirected networks to enable more robust community detection in weighted and directed networks. We derive a weighted and directed prior network that can incorporate metadata information and show how an efficient implementation in the community-detection method Infomap provides more reliable communities even with a significant fraction of data missing.

cs.SI↗

arXiv2021

Metadata-informed community detection with lazy encoding using absorbing random walks

Integrating structural information and metadata, such as gender, social status, or interests, enriches networks and enables a better understanding of the large-scale structure of complex systems. However, existing approaches to metadata integration only consider immediately adjacent nodes, thus failing to identify and exploit long-range correlations between metadata and network structure, typical of many spatial and social systems. Here we show how a flow-based community-detection approach can integrate network information and distant metadata, providing a more nuanced picture of network structure and correlations. We analyse social and spatial networks using the map equation framework and find that our methodology can detect a variety of useful metadata-informed partitions in diverse real-world systems. This framework paves the way for systematically incorporating metadata in network analysis.

physics.soc-ph↗