SCSean P. Cornelius

Sean P. Cornelius

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Sean P. Cornelius has a total of 38 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • How more data can hurt: Instability and regularization in next-generation reservoir computing

    by Yuanzhao Zhang, Edmilson Roque dos Santos, Huixin Zhang, Sean P. Cornelius

    It has been found recently that more data can, counter-intuitively, hurt the performance of deep neural networks. Here, we show that a more extreme version of the phenomenon occurs in data-driven models of dynamical systems. To elucidate the underlying mechanism, we focus on next-generation reservoir computing (NGRC)—a popular framework for learning dynamics from data. We find that, despite learning a better representation of the flow map with more training data, NGRC can adopt an ill-conditioned “integrator” and lose stability. We link this data-induced instability to the auxiliary dimensions created by the delayed states in NGRC. Based on these findings, we propose simple strategies to mitigate the instability, either by increasing regularization strength in tandem with data size, or by carefully introducing noise during training. Our results highlight the importance of proper regularization in data-driven modeling of dynamical systems.


  • Open Access

    SEEDNet: Covariate-free multi-country settlement-level epidemiological estimates datasets for network analysis

    by Amir Hossein Darooneh, Jean-Luc Kortenaar, Céline M. Goulart, Katie McLaughlin, Sean P. Cornelius, Diego G. Bassani


  • Open Access

    SEEDNet: Covariate-free multi-country settlement-level epidemiological estimates datasets for network analysis

    by Amir Hossein Darooneh, Jean-Luc Kortenaar, Céline M. Goulart, Katie McLaughlin, Sean P. Cornelius, Diego G. Bassani

    The study of population health through network science holds high promise, but data sources that allow complete representation of populations are limited in low-and middle-income countries. Large national health surveys designed to gather nationally representative health and development data are promising data sources but are not designed to produce small-area estimates of health indicators. Methods for producing these from national surveys tend to rely on varied covariate data sources and are computationally demanding, limiting their use for network representations of populations. To reduce the sources of measurement error and allow efficient multi-country representation of populations as networks of human settlements here, we present SEEDNet (Settlement-level Epidemiological Estimates Datasets for Network Analysis) 1 , a data library of multi-country representations of population health across human settlements. Our covariate-free method uses georeferenced national surveys to produce SAEs of health indicators through local inverse-distance weighted interpolation and includes an algorithm for the comprehensive identification of population settlements of all sizes across the globe. Our estimates are cross-validated against those obtained using a Bayesian Geostatistical Model. The method is fully automated, requiring a single standard georeferenced survey data source for mapping populations, eliminating the need for indicator or country-specific covariate selection by investigators. Computational efficiency is achieved by restricting computation to human-occupied areas and by adopting a logical aggregation of estimates into the complete range of settlement sizes. Standardized georeferenced settlement-level datasets for 15 indicators and 10 countries were validated in this paper, as well as the novel method to identify settlements. SEEDNet 1 is a specialized library of nodes that can serve as a basis for network representations of population health in low-and middle-income countries.

  • Open Access

    Mitochondrial Network Branching Enables Rapid Protein Spread with Slower Mitochondrial Dynamics

    by Prabha Chuphal, Jordan D. Lanctôt, Sean P. Cornelius, Aidan I. Brown

    Mitochondrial network structure is controlled by the dynamical processes of fusion and fission, which merge and split mitochondrial tubes into structures including branches and loops. To investigate the impact of mitochondrial network dynamics and structure on the spread of proteins and other molecules through mitochondrial networks, we used stochastic simulations of two distinct quantitative models that each included mitochondrial fusion and fission, and particle diffusion via the network. Better-connected mitochondrial networks and networks with faster dynamics exhibit more rapid particle spread on the network, with little further improvement once a network has become well connected. As fragmented networks gradually become better connected, particle spread either steadily improves until the networks become well connected for slow-diffusing particles or plateaus for fast-diffusing particles. We compared model mitochondrial networks with both end-to-end and end-to-side fusion, which form branches, to nonbranching model networks that lack end-to-side fusion. To achieve the optimum (most rapid) spread that occurs on well-connected branching networks, nonbranching networks require much faster fusion and fission dynamics. Thus, the process of end-to-side fusion, which creates branches in mitochondrial networks, enables rapid spread of particles on the network with relatively slow fusion and fission dynamics. This modeling of protein spread on mitochondrial networks builds toward mechanistic understanding of how mitochondrial structure and dynamics regulate mitochondrial function. Published by the American Physical Society 2024


  • Open Access

    Mastering percolation-like games with deep learning

    by Michael M. Danziger, Omkar R. Gojala, Sean P. Cornelius

    Though robustness of networks to random attacks has been widely studied, intentional destruction by an intelligent agent is not tractable with previous methods. Here we devise a single-player game on a lattice that mimics the logic of an attacker attempting to destroy a network. The objective of the game is to disable all nodes in the fewest number of steps. We develop a reinforcement learning approach using deep Q-learning that is capable of learning to play this game successfully, and in so doing, to optimally attack a network. Because the learning algorithm is universal, we train agents on different definitions of robustness and compare the learned strategies. We find that superficially similar definitions of robustness induce different strategies in the trained agent, implying that optimally attacking or defending a network is sensitive to the particular objective. Our method provides an approach to understand network robustness, with potential applications to other discrete processes in disordered systems.

  • Open Access

    Catch-22s of reservoir computing

    by Yuanzhao Zhang, Sean P. Cornelius

    Reservoir computing (RC) is a simple and efficient model-free framework for forecasting the behavior of nonlinear dynamical systems from data. Here, we show that there exist commonly-studied systems for which leading RC frameworks struggle to learn the dynamics unless key information about the underlying system is already known. We focus on the important problem of basin prediction—determining which attractor a system will converge to from its initial conditions. First, we show that the predictions of standard RC models (echo state networks) depend critically on warm-up time, requiring a warm-up trajectory containing almost the entire transient in order to identify the correct attractor. Accordingly, we turn to next-generation reservoir computing (NGRC), an attractive variant of RC that requires negligible warm-up time. By incorporating the exact nonlinearities in the original equations, we show that NGRC can accurately reconstruct intricate and high-dimensional basins of attraction, even with sparse training data (e.g., a single transient trajectory). Yet, a tiny uncertainty in the exact nonlinearity can render prediction accuracy no better than chance. Our results highlight the challenges faced by data-driven methods in learning the dynamics of multistable systems and suggest potential avenues to make these approaches more robust.


  • Top Cited Papers
    Open Access

    Evolution of cooperation on temporal networks

    by Aming Li, Lei Zhou, Qi Su, Sean P. Cornelius, Yang-Yu Liu, Long Wang, Simon A. Levin

    Population structure is a key determinant in fostering cooperation among naturally self-interested individuals in microbial populations, social insect groups, and human societies. Traditional research has focused on static structures, and yet most real interactions are finite in duration and changing in time, forming a temporal network. This raises the question of whether cooperation can emerge and persist despite an intrinsically fragmented population structure. Here we develop a framework to study the evolution of cooperation on temporal networks. Surprisingly, we find that network temporality actually enhances the evolution of cooperation relative to comparable static networks, despite the fact that bursty interaction patterns generally impede cooperation. We resolve this tension by proposing a measure to quantify the amount of temporality in a network, revealing an intermediate level that maximally boosts cooperation. Our results open a new avenue for investigating the evolution of cooperation and other emergent behaviours in more realistic structured populations.


  • Who holds the power? The Square and the Tower: Networks and Power, from the Freemasons to Facebook Niall Ferguson Penguin Press, 2018. 607 pp.

    by Sean P. Cornelius

    What was the cause of Donald Trump9s stunning victory over Hillary Clinton in the 2016 U.S. presidential election? Was it the peculiarities of the electoral college? Voter resistance to three-term rule by a single party? Anxiety about illegal immigration? As Niall Ferguson explains in The Square and the Tower, the answer lies largely in one word: networks.

  • Top Cited Papers

    The fundamental advantages of temporal networks

    by A. Li, S. P. Cornelius, Y.-Y. Liu, L. Wang, A.-L. Barabási

    In network science, change is good: Historically, network science focused on static networks, in which nodes are connected by permanent links. However, in networked systems ranging from protein-protein interactions to social networks, links change. Although it might seem that permanent links would make it easier to control a system, Li et al. demonstrate that temporality has advantages in real and simulated networks. Temporal networks can be controlled more efficiently and require less energy than their static counterparts. Science , this issue p. 1042