publications | Guillermo Romero Moreno

preprints & under review

Pitfalls in understanding how multiple long-term conditions cluster: whole population and age-stratified associations in 7,490,874 people in England

Guillermo Romero Moreno, Valerio Restocchi, Nazir Lone, Jacques D. Fleuriot, Jake Palmer, Luna De Ferrari, and Bruce Guthrie

medRxiv, Feb 2025

Preprint. Under review at Nature Communications

Abs DOI Bib HTML

Studies of how multiple long-term conditions (MLTC) cluster together in individuals vary in the populations studied, and whether they age and/or sex stratify, which limits comparison between studies and reproducibility. This study uses a large, UK primary-care dataset to examine how pairwise strength of association between 74 conditions varies by age in both men and women aged 30-99 years, and to explore implications for MLT cluster analyses. Joint prevalence of conditions was lowest in younger age-groups and progressively increased with age, whereas Association Beyond Chance (ABC) was highest in younger age-groups and progressively decreased with age. Condition clustering based on ABC identified different clusters in all men and all women aged 30-99 years, and these clusters differed from those identified in each age-group. Researchers examining how MLTC cluster should consider whether age and sex stratification is appropriate given their study aims and/or would improve comparability and reproducibility, and explicitly justify their choices.
@article{romero_moreno_pitfalls_2025, title = {Pitfalls in understanding how multiple long-term conditions cluster: whole population and age-stratified associations in 7,490,874 people in {England}}, copyright = {All rights reserved}, journal = {medRxiv}, note = {Preprint. Under review at Nature Communications}, url = {https://www.medrxiv.org/content/10.1101/2025.02.06.25321779v1}, doi = {10.1101/2025.02.06.25321779}, author = {Romero Moreno, Guillermo and Restocchi, Valerio and Lone, Nazir and Fleuriot, Jacques D. and Palmer, Jake and De Ferrari, Luna and Guthrie, Bruce}, month = feb, year = {2025}, }
Towards reproducible multimorbidity clustering in electronic health records: a transparent pipeline for aligning research aims and methodology

Guillermo Romero Moreno, Valerio Restocchi, Luna De Ferrari, Jake Palmer, Jacques Fleuriot, Bruce Guthrie, and Nazir Lone

medRxiv, May 2026

Preprint.

Abs DOI Bib HTML

The availability of electronic health records has facilitated data-driven approaches to the understanding of multimorbidity, with clustering becoming a common tool for uncovering relevant groups of associated conditions. Previous studies, however, have found challenges in their reproducibility, with wide disparity in the reported clusters. At the core of this issue lays a vagueness of the definition of a cluster, leading to a lack of standards in their methods and evaluation, while implementation details are often not completely reported or explicit in their assumptions. We present a methodological pipeline that can be adapted to different cluster definitions (e.g. multiple cluster membership or clusters where all nodes are mutually associated) and a set of scores that can be composed into an evaluation metric that explicitly incorporates assumptions that align with the research aims. We apply our pipeline to a healthcare dataset of over 7 million patients in England and show how clusters may drastically differ when varying the parameter choices, exposing the risks of reporting a single clustering realisation. Our methodological pipeline, evaluation framework, and tools for analysis and network visualisation serve as a reference to transparently explore and align methodological decisions to the aims of multimorbidity clustering, contributing to overcome the reproducibility challenges of the field.
@article{romero_moreno_reproducible_2026, doi = {10.64898/2026.05.25.26353178}, title = {Towards reproducible multimorbidity clustering in electronic health records: a transparent pipeline for aligning research aims and methodology}, copyright = {All rights reserved}, journal = {medRxiv}, note = {Preprint.}, url = {https://www.medrxiv.org/content/10.64898/2026.05.25.26353178v1}, author = {Romero Moreno, Guillermo and Restocchi, Valerio and De Ferrari, Luna and Palmer, Jake and Fleuriot, Jacques and Guthrie, Bruce and Lone, Nazir}, month = may, year = {2026}, }
Modelling the complex dynamics of online disinformation cascades

Guillermo Romero Moreno, Valentin Robu, and Han La Poutré

arXiv preprint arXiv:2610.13839, Oct 2026

Accepted for The 15th International Conference on Complex Networks and their Applications

Abs DOI arXiv Bib HTML PDF

Disinformation on online social media has been on the rise in the past years. While there is extensive literature modelling social cascades, online disinformation cascades show distinctive characteristics: they provoke debunking reactions, they spread over networks with heterogeneous connectivity, and they compete for the limited attention of users — which recommendation algorithms may partly offset. We propose a model including these ingredients and study it through Monte–Carlo simulations and mean–field approximations, on random graphs and an empirical Twitter network. We find that debunking reactions are double–edged: they reduce the reach of appealing items but allow unappealing ones to reach audiences they would not otherwise reach. Degree heterogeneity makes cascades trigger at a much lower content appeal but limits their final reach: heterogeneous networks are easy to ignite yet hard to saturate. Limited attention cancels the effects from degree heterogeneity, although the mediation of recommendation algorithms may restore them.
@article{romero_moreno_disinformation_2026, doi = {10.48550/arXiv.2610.13839}, title = {Modelling the complex dynamics of online disinformation cascades}, copyright = {All rights reserved}, journal = {arXiv preprint arXiv:2610.13839}, note = {Accepted for The 15th International Conference on Complex Networks and their Applications}, url = {https://arxiv.org/abs/2610.13839}, author = {Romero Moreno, Guillermo and Robu, Valentin and La Poutr{\'e}, Han}, month = oct, year = {2026}, }
SimPol: Simulating polarisation in political belief networks in European countries

Isabela Burattini Freire, Hongryol Cha, Irina Epure, Sara Filippini, Karan K. H. Manjunatha, Chethan Kavaraganahalli Prasanna, Ivan Samoylenko, Niels Van Santen, Adarsh Prabhakaran, and Guillermo Romero Moreno

arXiv preprint arXiv:2606.27968, Jun 2026

Preprint. Complexity72h 2026 project

Abs DOI arXiv HTML PDF

Here we combine empirical network analysis with agent-based modelling to understand how different ways of structuring belief systems may affect the polarisation drive, and how the diversity of belief systems in Europe may result in different polarisation trajectories. Using the 2016 European Social Survey, we infer belief networks across 23 European countries via a Bayesian algorithm, revealing that belief systems are predominantly organised around immigration, LGBT rights, and economic interventionism, reflecting the influence of populist discourse across the continent. We further verify a Western-Eastern divide across the national belief networks: in Western European countries, left-right self-identification is a more reliable predictor of broader belief alignment, whereas in Eastern Europe this relationship breaks down. By applying these empirical belief networks into a sociologically grounded agent-based model, we further show that polarisation is amplified by high individual belief rigidity and low susceptibility to social influence, and that cross-country differences in polarisation levels mirror the same geographic divide observed in belief network topology. These findings establish belief networks topologies as a structural driver of political polarisation, with implications for understanding and anticipating polarisation dynamics across diverse European contexts. We find that populations are not polarised when little attention is placed on maintaining internal coherence and polarisation levels are moderate when high attention is placed in both keeping internal coherence and agreement in beliefs with others.

published

2025

Analysing Opinion Dynamics via a Cognitive Model of Structured Beliefs

E. Michaelidou, E. Ioannou, K. Tapper, B. D. Goddard, and Guillermo Romero Moreno

In Artificial Life and Evolutionary Computation. WIVACE 2024, 2025

Abs DOI Bib

Opinion dynamics models can help explain observed social phenomena that emerge from social interaction, such as echo chambers or polarisation, by using computational and numerical methods. However, to make them mathematically tractable, many such models are based on simplistic and unrealistic assumptions that do not align with the knowledge from the empirical cognitive science field. A significant exception is a recent multidimensional opinion model proposed by Mueller and Tan, which was able to reproduce opinion divergence and group polarisation by establishing a need for consistency between related beliefs that restricts the space of allowed combinations of opinions. Although this contribution is promising in closing the gap between the cognitive sciences and opinion dynamics models, the authors only provided a limited set of experiments to show the behaviour of the model and a more in-depth understanding is lacking. Here, we provide a concise mathematical analysis and a more thorough numerical analysis of the cognitive model of opinion dynamics proposed by Mueller and Tan and show that the outcomes of opinion divergence and group polarisation only emerge under some regimes of the parameter space: medium-term dynamics, a moderate number of topics discussed upon interaction, with opinion divergence emerging under sparse belief spaces and group polarisation under large, dense belief spaces. These scenarios should be contrasted with real-life cases to confirm whether the assumptions of the constrained belief space are enough to explain the observed phenomena or whether additional ingredients should be added for a more realistic representation of social behaviour.
@inproceedings{michaelidou_analysing_2025, title = {Analysing {Opinion} {Dynamics} via a {Cognitive} {Model} of {Structured} {Beliefs}}, copyright = {All rights reserved}, booktitle = {Artificial {Life} and {Evolutionary} {Computation}. {WIVACE} 2024}, editor = {Carletti, T. and Njougouo, T. S. and Tuci, E.}, series = {Communications in {Computer} and {Information} {Science}}, volume = {2532}, publisher = {Springer, Cham}, doi = {10.1007/978-3-031-93631-9_3}, author = {Michaelidou, E. and Ioannou, E. and Tapper, K. and Goddard, B. D. and Romero Moreno, Guillermo}, year = {2025}, pages = {28--41}, }

2024

Influence Maximisation and Adaptive Control of Opinion Dynamics on Complex Networks

Guillermo Romero Moreno

University of Southampton, 2024

Abs Bib HTML

Opinion shaping, by which a strategic attempts to influence the opinion of a social group, is a pervasive phenomenon in human behaviour, with clear examples in current societies in information campaigns, political competition, or marketing. However, the effect that influence attempts have in societies is not easy to foresee, as they are complex systems where interactions can cascade and compound in unpredictable ways. The research subfield of opinion dynamics employs a mathematical modelling approach to study these phenomena using tools from the more general field of complex systems. In this approach, the opinions held by an individual are typically modelled as mathematical variables, with changes in opinion dictated by simple rules triggering upon interactions, which typically only happen interact according to a complex network that reflects their social structure. By drawing from techniques from statistical physics, agent–based simulations, and optimisation research, this thesis studies the effects that different external control strategies have on opinion formation processes. We first focus on a case where the external controller is a ‘perfect optimiser’, i.e. they can split their influence targets among the individuals and have information to do so strategically as to achieve the best possible result — a problem commonly known as Influence Maximisation. We place this optimising controller against an opponent in a scenario where individuals have preferences over two possible choices (e.g. products or political parties), propose an optimisation algorithm to find optimal targetting strategies, and analyse the characteristics of these to understand why they are effective. We find that optimal strategies can be characterised by two heuristics, shadowing and shielding, while the structure of the network only plays a secondary role. Shadowing entails targeting the same individuals as the opponent to directly block her influence in the network, unless the opponent’s influencing power is much higher, in which case these are avoided. Shielding entails ring–fencing the individuals targeted by the opponent to indirectly block the spreading of her influence. We then modify the previous scenario to incorporate different levels of bias against either opinion in individuals, and analyse optimal targetting strategies when the population is structured in different network topologies. We find a general pattern in which individuals that are difficult to control (i.e. biased against the controller, highly connected, or targeted by the opponent) are avoided if the influencing power is small and sought if the influencing power is high. Last, we shift the scenario to one in which the controllers are not ‘perfect optimisers’ any more but only have very limited information and perform local moves to improve their situation in the short term. Therefore, their control strategies are ‘adaptive’, reacting to the dynamics of the opinions as they unfold and creating a dynamic interaction between the two. We focus on the specific agenda–setting scenario where political parties seek to increase votes by affecting the importance that different political dimensions have and how their strategies interferes with the process of arriving at consensus or polarisation within the social group. We find in this scenario that party competition often fosters the arrival of a polarised state with most individuals gathering in two opposed camps, although if parties perform frequent shifts in the issues they give importance to, their behaviour inadvertently fosters the arrival at a consensus in opinion in social group.
@phdthesis{romero_moreno_influence_2024, address = {Southampton}, title = {Influence {Maximisation} and {Adaptive} {Control} of {Opinion} {Dynamics} on {Complex} {Networks}}, copyright = {All rights reserved}, url = {https://eprints.soton.ac.uk/485922/}, language = {English}, school = {University of Southampton}, author = {Romero Moreno, Guillermo}, year = {2024}, }
Multimorbidity analysis with low condition counts: a robust Bayesian approach for small but important subgroups

Guillermo Romero Moreno, Valerio Restocchi, Jacques D. Fleuriot, Atul Anand, Stewart W. Mercer, and Bruce Guthrie

eBioMedicine, Apr 2024

Abs DOI Bib

Background Robustly examining associations between long-term conditions may be important in identifying opportunities for intervention in multimorbidity but is challenging when evidence is limited. We have developed a Bayesian inference framework that is robust to sparse data and used it to quantify morbidity associations in the oldest old, a population with limited available data. Methods We conducted a retrospective cross-sectional study of a representative dataset of primary care patients in Scotland as of March 2007. We included 40 long-term conditions and studied their associations in 12,009 individuals aged 90 and older, stratified by sex (3039 men, 8970 women). We analysed associations obtained with Relative Risk (RR), a standard measure in the literature, and compared them with our proposed measure, Associations Beyond Chance (ABC). To enable a broad exploration of interactions between long-term conditions, we built networks of association and assessed differences in their analysis when associations are estimated by RR or ABC. Findings Our Bayesian framework was appropriately more cautious in attributing association when evidence is lacking, particularly in uncommon conditions. This caution in reporting association was also present in reporting differences in associations between sex and affected the aggregated measures of multimorbidity and network representations. Interpretation Incorporating uncertainty into multimorbidity research is crucial to avoid misleading findings when evidence is limited, a problem that particularly affects small but important subgroups. Our proposed framework improves the reliability of estimations of associations and, more in general, of research into disease mechanisms and multimorbidity. Funding National Institute for Health and Care Research.
@article{romero_moreno_multimorbidity_2024, title = {Multimorbidity analysis with low condition counts: a robust {Bayesian} approach for small but important subgroups}, volume = {102}, copyright = {All rights reserved}, issn = {2352-3964}, shorttitle = {Multimorbidity analysis with low condition counts}, url = {https://www.sciencedirect.com/science/article/pii/S2352396424001166}, doi = {10.1016/j.ebiom.2024.105081}, urldate = {2024-07-10}, journal = {eBioMedicine}, author = {Romero Moreno, Guillermo and Restocchi, Valerio and Fleuriot, Jacques D. and Anand, Atul and Mercer, Stewart W. and Guthrie, Bruce}, month = apr, year = {2024}, keywords = {Network analysis, Multimorbidity, Bayesian inference, Relative risk, Association measures, Low counts}, pages = {105081}, }

2022

Sensing Enhancement on Social Networks: The Role of Network Topology

Markus Brede, and Guillermo Romero Moreno

Entropy, May 2022

Abs DOI Bib

Sensing and processing information from dynamically changing environments is essential for the survival of animal collectives and the functioning of human society. In this context, previous work has shown that communication between networked agents with some preference towards adopting the majority opinion can enhance the quality of error-prone individual sensing from dynamic environments. In this paper, we compare the potential of different types of complex networks for such sensing enhancement. Numerical simulations on complex networks are complemented by a mean-field approach for limited connectivity that captures essential trends in dependencies. Our results show that, whilst bestowing advantages on a small group of agents, degree heterogeneity tends to impede overall sensing enhancement. In contrast, clustering and spatial structure play a more nuanced role depending on overall connectivity. We find that ring graphs exhibit superior enhancement for large connectivity and that random graphs outperform for small connectivity. Further exploring the role of clustering and path lengths in small-world models, we find that sensing enhancement tends to be boosted in the small-world regime.
@article{Brede2022, title = {Sensing {Enhancement} on {Social} {Networks}: {The} {Role} of {Network} {Topology}}, volume = {24}, copyright = {All rights reserved}, issn = {1099-4300}, url = {https://www.mdpi.com/1099-4300/24/5/738}, doi = {10.3390/e24050738}, number = {5}, journal = {Entropy}, author = {Brede, Markus and Romero Moreno, Guillermo}, month = may, year = {2022}, keywords = {Opinion dynamics, Complex networks, Sensing enhancement}, pages = {738}, }
Sensing Enhancement on Complex Networks

Markus Brede, and Guillermo Romero Moreno

In Complex Networks & Their Applications X, May 2022

Abs DOI Bib

Previous work has shown that communication between agents with some preference towards adopting the majority opinion can enhance the quality of error-prone individual sensing from dynamic environments. In this paper, we compare the potential of different types of complex networks for such sensing enhancement. Numerical simulations on complex networks are complemented by a mean-field approach for limited connectivity that captures essential trends in dependencies. Our results show that whilst bestowing advantages on a small group of agents degree heterogeneity tends to impede overall sensing enhancement, while clustering and spatial structure play a more nuanced role depending on overall connectivity. We find that for low connectivity sensing enhancement is maximised by random regular networks, whereas for large connectivity best sensing enhancement is found for ring graphs.
@inproceedings{brede_sensing_2021, title = {Sensing {Enhancement} on {Complex} {Networks}}, copyright = {All rights reserved}, booktitle = {Complex {Networks} \& {Their} {Applications} {X}}, series = {Studies in {Computational} {Intelligence}}, volume = {1073}, publisher = {Springer, Cham}, isbn = {978-3-030-93412-5}, doi = {10.1007/978-3-030-93413-2_30}, author = {Brede, Markus and Romero Moreno, Guillermo}, year = {2022}, pages = {353--364}, }
Learning VAA: A new method for matching users to parties in voting advice applications

Guillermo Romero Moreno, Javier Padilla, and Enrique Chueca

Journal of Elections, Public Opinion and Parties, May 2022

Abs DOI Bib

Voting Advice Applications (VAAs) aim to increase voters’ political competence by providing them with the closest political party according to their preferences. To do this, VAAs usually compare and aggregate the positions of users and political parties on a set of policy issues by defining a conceptual space and some distance metric on it. In this paper, we argue that the main method for performing the comparison adapts to users’ preferences unsatisfactorily because 1) they use unjustified a priori decisions for weighting policy issues and 2) they employ the same issue-voting space on all policy issues. Some exceptional cases address these issues by providing a community-based recommendation, but often come with lack of interpretability. To fill these gaps, we propose an adaptive algorithm that learns the configuration of the conceptual space from users’ answers. We employ a hybrid VAA that uses expert coding for the party positions and users’ data to adjust the calculation of the distance between users and parties. This new matching method, the Learning VAA, innovates by adjusting the saliency and issue-voting space for every policy issue. We argue and empirically demonstrate that our model fits better the users’ preferences while providing a higher degree of interpretability.
@article{romero_moreno_learning_2022, title = {Learning {VAA}: {A} new method for matching users to parties in voting advice applications}, volume = {32}, copyright = {All rights reserved}, url = {https://www.tandfonline.com/doi/full/10.1080/17457289.2020.1760282}, doi = {10.1080/17457289.2020.1760282}, number = {2}, journal = {Journal of Elections, Public Opinion and Parties}, author = {Romero Moreno, Guillermo and Padilla, Javier and Chueca, Enrique}, year = {2022}, pages = {339--357}, }
Building co-morbidity networks via Bayesian network reconstruction

Guillermo Romero Moreno, Valerio Restocchi, and Jacques D. Fleuriot

In The 11th International Conference on Complex Networks and their Applications: Book of Abstracts, Nov 2022

Abs Bib HTML PDF

Patients that simultaneously suffer multiple long-term health conditions pose a problem to current healthcare systems, as these are configured for individual conditions and overlook their interaction. A particularly useful approach to represent co-morbidity data is via networks, where nodes correspond to conditions and edges represent their relations. However, determining the degree of co-morbidity and deciding on their relevance is not trivial. Here, we adapt a Bayesian network reconstruction method developed by Young et al. to tackle the problems of co-morbidity network construction. This methodology allows us to infer the network of relevant associations between conditions without any bias or need to choose an arbitrary threshold. Instead, relevant associations are determined via a statistically sound method that accounts for the noise and uncertainty in the data. Furthermore, the Bayesian framework retrieves the whole posterior probability distribution, allowing us to factor in, or analyse independently, different hypotheses of models that generated the data.
@inproceedings{romero_moreno_building_2022, title = {Building co-morbidity networks via {Bayesian} network reconstruction}, copyright = {All rights reserved}, booktitle = {The 11th {International} {Conference} on {Complex} {Networks} and their {Applications}: {Book} of {Abstracts}}, address = {Palermo, Italy}, author = {Romero Moreno, Guillermo and Restocchi, Valerio and Fleuriot, Jacques D.}, month = nov, year = {2022}, pages = {491--493}, }

2021

Shadowing and shielding: Effective heuristics for continuous influence maximisation in the voting dynamics

Guillermo Romero Moreno, Sukankana Chakraborty, and Markus Brede

PLOS ONE, Jun 2021

Abs DOI Bib

Influence maximisation, or how to affect the intrinsic opinion dynamics of a social group, is relevant for many applications, such as information campaigns, political competition, or marketing. Previous literature on influence maximisation has mostly explored discrete allocations of influence, i.e. optimally choosing a finite fixed number of nodes to target. Here, we study the generalised problem of continuous influence maximisation where nodes can be targeted with flexible intensity. We focus on optimal influence allocations against a passive opponent and compare the structure of the solutions in the continuous and discrete regimes. We find that, whereas hub allocations play a central role in explaining optimal allocations in the discrete regime, their explanatory power is strongly reduced in the continuous regime. Instead, we find that optimal continuous strategies are very well described by two other patterns: (i) targeting the same nodes as the opponent (shadowing) and (ii) targeting direct neighbours of the opponent (shielding). Finally, we investigate the game-theoretic scenario of two active opponents and show that the unique pure Nash equilibrium is to target all nodes equally. These results expose fundamental differences in the solutions to discrete and continuous regimes and provide novel effective heuristics for continuous influence maximisation.
@article{RomeroMoreno2021, title = {Shadowing and shielding: {Effective} heuristics for continuous influence maximisation in the voting dynamics}, volume = {16}, copyright = {All rights reserved}, issn = {1932-6203}, url = {https://dx.plos.org/10.1371/journal.pone.0252515}, doi = {10.1371/journal.pone.0252515}, number = {6}, journal = {PLOS ONE}, author = {Romero Moreno, Guillermo and Chakraborty, Sukankana and Brede, Markus}, editor = {Cherifi, Hocine}, month = jun, year = {2021}, keywords = {Influence Maximisation, Opinion Dynamics, Voter model, Voter Model, Social network, Influence maximization, opinion dynamics, complex networks}, pages = {e0252515}, }
A normative set of criteria to increase political competence through voting advice applications

Javier Padilla, Javier Ramos, Guillermo Romero Moreno, and Jesús Enrique Chueca

International Journal of Electronic Governance, Jul 2021

Abs DOI Bib

This paper argues that current voting advice applications (VAAs) do not sufficiently fulfil their stated aim of increasing voters’ political competence. First, we define four criteria to evaluate whether their methods are likely to increase political competence: informativeness, respect for users’ way of comparing and aggregating policy issues, reliability, and transparency. Second, we argue that current VAAs compare and aggregate users’ and parties’ policy preferences following a weak method that fails in two of them. To prove it, we analyse the methodology of most currents VAAs and use the outcomes from the EU-Vox 2014 in several countries. Third, we discuss two possibilities by which VAAs could improve: (1) by using ex-ante survey data to fill their gaps, or (2) by creating a learning algorithm to adapt the VAA to users’ preferences. We found that some changes need to be made if VAAs aim to have an impact on users’ political competence.
@article{Padilla2021, title = {A normative set of criteria to increase political competence through voting advice applications}, volume = {13}, copyright = {All rights reserved}, issn = {1742-7509}, url = {http://www.inderscience.com/link.php?id=116869}, doi = {10.1504/IJEG.2021.116869}, number = {2}, journal = {International Journal of Electronic Governance}, author = {Padilla, Javier and Ramos, Javier and Romero Moreno, Guillermo and Chueca, Jesús Enrique}, month = jul, year = {2021}, pages = {149--175}, }
The effects of party competition on consensus formation

Guillermo Romero Moreno, Javier Padilla, and Markus Brede

In 3rd International Workshop on Agent-Based Modelling of Human Behaviour, Jul 2021

Abs Bib HTML

The fight over setting the political agenda is one of the basic mechanisms of party competition of every democracy. How- ever, this political game may have side effects in other as- pects of the public debate. One aspect of general interest is how it may alter consensus formation processes among citi- zens, which may result in states of consensus, polarisation, or opinion fragmentation in the population. In this paper, we study the interrelated dynamics of two processes affecting opinion dynamics when multiple issues are debated. First, we model party competition via campaigning and its effect on the saliency or importance with which citizens perceive different political issues. Second, we consider a bounded–confidence model to describe the dynamics of citizens’ opinion and con- sensus formation. We find that the effects of party competi- tion on consensus formation are rich and non-trivially depen- dent on the configuration of party positions in the political space. We illustrate that —as one would intuitively expect— there are party configurations that foster a paradigmatic state of polarisation for a wide range of model parameters. How- ever, we also show that other party configurations have the opposite effect, and can facilitate reaching a consensus state that could otherwise not have been achieved. Our results il- lustrate the richness of possible outcomes of interrelations be- tween party competition and consensus formation. Introduction
@inproceedings{romero_moreno_effects_2021, title = {The effects of party competition on consensus formation}, copyright = {All rights reserved}, booktitle = {3rd {International} {Workshop} on {Agent}-{Based} {Modelling} of {Human} {Behaviour}}, url = {https://eprints.soton.ac.uk/449484/}, author = {Romero Moreno, Guillermo and Padilla, Javier and Brede, Markus}, year = {2021}, }

2020

Continuous Influence Maximisation for the Voter Dynamics: Is Targeting High-Degree Nodes a Good Strategy? Extended Abstract

Guillermo Romero Moreno, Long Tran-Thanh, and Markus Brede

In Proceedings of the 19th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2020), May 2020

Abs DOI Bib

In this paper, we relate influence maximisation (IM) for the voting dynamics to models of network control in which external controllers interact with the intrinsic dynamics of opinion spread. In contrast to previous literature, which has mostly explored the discrete setting, our focus is on continuous allocations of control. We develop an algorithm to numerically solve our IM problem via gradient ascent. We explore optimal allocations for leader-follower type networks for different budget scenarios and observe that optimal allocations do not systematically target hub nodes, as it has been found in previous literature. Conversely, strategies are strongly opponent-depend, avoiding nodes targeted by the opponent if the opponent has a larger budget, while shadowing the opponent’s allocation otherwise, i.e. targeting the same nodes as them.
@inproceedings{RomeroMoreno2020a, doi = {10.65109/jntx8567}, title = {Continuous {Influence} {Maximisation} for the {Voter} {Dynamics}: {Is} {Targeting} {High}-{Degree} {Nodes} a {Good} {Strategy}? {Extended} {Abstract}}, copyright = {All rights reserved}, booktitle = {Proceedings of the 19th {International} {Conference} on {Autonomous} {Agents} and {Multiagent} {Systems} ({AAMAS} 2020)}, address = {Auckland, New Zealand}, url = {https://eprints.soton.ac.uk/438158/}, author = {Romero Moreno, Guillermo and Tran-Thanh, Long and Brede, Markus}, month = may, year = {2020}, keywords = {Influence Maximisation, Voter Model, Zealots, influence maximization, complex networks, voter model, external, external control}, pages = {1981--1983}, }
Zealotry and Influence Maximization in the Voter Model: When to Target Partial Zealots?

Guillermo Romero Moreno, Edoardo Manino, Long Tran-Thanh, and Markus Brede

In Complex Networks XI: Proceedings of the 11th Conference on Complex Networks CompleNet 2020, May 2020

ISSN: 22138692

Abs DOI arXiv Bib

In this paper, we study influence maximization in the voter model in the presence of biased voters (partial zealots) on complex networks. Under what conditions should an external controller with finite budget who aims at maximizing its influence over the system target partial zealots? Our analysis, based on both analytical and numerical results, shows a rich diagram of preferences and degree-dependencies of allocations to partial zealots and normal agents varying with the budget. We find that when we have a large budget or for low levels of zealotry, optimal strategies should give larger allocations to partial zealots and allocations are positively correlated with node degree. In contrast, for low budgets or highly-biased zealots, optimal strategies give higher allocations to normal agents, with some residual allocations to partial zealots, and allocations to both types of agents decrease with node degree. Our results emphasize that heterogeneity in agent properties strongly affects strategies for influence maximization on heterogeneous networks.
@incollection{RomeroMoreno2020, title = {Zealotry and {Influence} {Maximization} in the {Voter} {Model}: {When} to {Target} {Partial} {Zealots}?}, copyright = {All rights reserved}, isbn = {978-3-030-40942-5}, url = {http://link.springer.com/10.1007/978-3-030-40943-2_10}, booktitle = {Complex {Networks} {XI}: {Proceedings} of the 11th {Conference} on {Complex} {Networks} {CompleNet} 2020}, series = {Springer {Proceedings} in {Complexity}}, publisher = {Springer, Cham}, author = {Romero Moreno, Guillermo and Manino, Edoardo and Tran-Thanh, Long and Brede, Markus}, year = {2020}, doi = {10.1007/978-3-030-40943-2_10}, note = {ISSN: 22138692}, keywords = {Voter model, Zealots, Influence maximization, Complex networks}, pages = {107--118}, }

2019

Saliency Map on CNNs for Protein Secondary Structure Prediction

Guillermo Romero Moreno, Mahesan Niranjan, and Adam Prugel-Bennett

In ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 2019

Abs DOI Bib

Deep learning, a powerful methodology for data-driven modelling, has been shown to be useful in tackling several problems in the biomedical domain. However, deep neural architectures lack interpretability of how predictions from them are made on any test input. While several approaches to “opening the black box” are being developed, their application to biological and medical data is very much as its infancy. Here, we consider the specific problem of protein secondary structure prediction using the techniques of saliency maps to explain decisions of a deep neural network. The analysis leads to two important observations: (a) one-hot-encoded amino-acids are irrelevant in the presence of PSSM values as extra features; and (b) in predicting α-helices at any position, amino-acids to the right are far more important than those to the left. The latter observation may have a biological basis relating to the synthesis of proteins by ribosome movement from left to right, sequentially adding amino-acids.
@inproceedings{romero_moreno_saliency_2019, title = {Saliency {Map} on {CNNs} for {Protein} {Secondary} {Structure} {Prediction}}, copyright = {All rights reserved}, booktitle = {{ICASSP} 2019 - 2019 {IEEE} {International} {Conference} on {Acoustics}, {Speech} and {Signal} {Processing} ({ICASSP})}, publisher = {IEEE}, doi = {10.1109/ICASSP.2019.8683603}, author = {Romero Moreno, Guillermo and Niranjan, Mahesan and Prugel-Bennett, Adam}, month = may, year = {2019}, pages = {1249--1253}, }
Shielding and Shadowing: A Tale of Two Strategies for Opinion Control in the Voting Dynamics

Guillermo Romero Moreno, Long Tran-Thanh, and Markus Brede

In Complex Networks and Their Applications VIII, Nov 2019

Abs DOI Bib

This paper focuses on influence maximization or opinion control in the voting dynamics on social networks. We show two simple heuristics that are effective strategies to enhance vote shares: (i) avoiding the nodes controlled by your opponent when having a lower budget while focusing on them when having a larger budget (shadowing) and (ii) ringfencing her influence by targeting control on adjacent nodes (shielding). The paper presents an empirical numerical evaluation of these strategies for various classes of complex networks which is backed up by analytical results obtained via a mean-field approach, in good agreement with numerical results. Importantly, we also show that optimal influence allocations tend to not be localized, but can include targeting nodes significant distances away from opposing influence.
@inproceedings{romero_moreno_shielding_2019, title = {Shielding and {Shadowing}: {A} {Tale} of {Two} {Strategies} for {Opinion} {Control} in the {Voting} {Dynamics}}, copyright = {All rights reserved}, booktitle = {Complex {Networks} and {Their} {Applications} {VIII}}, series = {Studies in {Computational} {Intelligence}}, volume = {881}, publisher = {Springer, Cham}, doi = {10.1007/978-3-030-36687-2_57}, author = {Romero Moreno, Guillermo and Tran-Thanh, Long and Brede, Markus}, month = nov, year = {2019}, pages = {682--693}, }