RESUMEN

Deep neural networks have made tremendous gains in emulating human-like intelligence, and have been used increasingly as ways of understanding how the brain may solve the complex computational problems on which this relies. However, these still fall short of, and therefore fail to provide insight into how the brain supports strong forms of generalization of which humans are capable. One such case is out-of-distribution (OOD) generalization - successful performance on test examples that lie outside the distribution of the training set. Here, we identify properties of processing in the brain that may contribute to this ability. We describe a two-part algorithm that draws on specific features of neural computation to achieve OOD generalization, and provide a proof of concept by evaluating performance on two challenging cognitive tasks. First we draw on the fact that the mammalian brain represents metric spaces using grid cell code (e.g., in the entorhinal cortex): abstract representations of relational structure, organized in recurring motifs that cover the representational space. Second, we propose an attentional mechanism that operates over the grid cell code using determinantal point process (DPP), that we call DPP attention (DPP-A) - a transformation that ensures maximum sparseness in the coverage of that space. We show that a loss function that combines standard task-optimized error with DPP-A can exploit the recurring motifs in the grid cell code, and can be integrated with common architectures to achieve strong OOD generalization performance on analogy and arithmetic tasks. This provides both an interpretation of how the grid cell code in the mammalian brain may contribute to generalization performance, and at the same time a potential means for improving such capabilities in artificial neural networks.

Asunto(s)

RESUMEN

Many techniques have been proposed to model space-varying observation processes with a nonstationary spatial covariance structure and/or anisotropy, usually on a geostatistical framework. Nevertheless, there is an increasing interest in point process applications, and methodologies that take nonstationarity into account are welcomed. In this sense, this work proposes an extension of a class of spatial Cox process using spatial deformation. The proposed method enables the deformation behavior to be data-driven, through a multivariate latent Gaussian process. Inference leads to intractable posterior distributions that are approximated via MCMC. The convergence of algorithms based on the Metropolis-Hastings steps proved to be slow, and the computational efficiency of the Bayesian updating scheme was improved by adopting Hamiltonian Monte Carlo (HMC) methods. Our proposal was also compared against an alternative anisotropic formulation. Studies based on synthetic data provided empirical evidence of the benefit brought by the adoption of nonstationarity through our anisotropic structure. A real data application was conducted on the spatial spread of the Spodoptera frugiperda pest in a corn-producing agricultural area in southern Brazil. Once again, the proposed method demonstrated its benefit over alternatives.

RESUMEN

Evaluation metrics for prediction error, model selection and model averaging on space-time data are understudied and poorly understood. The absence of independent replication makes prediction ambiguous as a concept and renders evaluation procedures developed for independent data inappropriate for most space-time prediction problems. Motivated by air pollution data collected during California wildfires in 2008, this manuscript attempts a formalization of the true prediction error associated with spatial interpolation. We investigate a variety of cross-validation (CV) procedures employing both simulations and case studies to provide insight into the nature of the estimand targeted by alternative data partition strategies. Consistent with recent best practice, we find that location-based cross-validation is appropriate for estimating spatial interpolation error as in our analysis of the California wildfire data. Interestingly, commonly held notions of bias-variance trade-off of CV fold size do not trivially apply to dependent data, and we recommend leave-one-location-out (LOLO) CV as the preferred prediction error metric for spatial interpolation.

RESUMEN

Objective. Oscillations figure prominently as neurological disease hallmarks and neuromodulation targets. To detect oscillations in a neuron's spiking, one might attempt to seek peaks in the spike train's power spectral density (PSD) which exceed a flat baseline. Yet for a non-oscillating neuron, the PSD is not flat: The recovery period ('RP', the post-spike drop in spike probability, starting with the refractory period) introduces global spectral distortion. An established 'shuffling' procedure corrects for RP distortion by removing the spectral component explained by the inter-spike interval (ISI) distribution. However, this procedure sacrifices oscillation-related information present in the ISIs, and therefore in the PSD. We asked whether point process models (PPMs) might achieve more selective RP distortion removal, thereby enabling improved oscillation detection.Approach. In a novel 'residuals' method, we first estimate the RP duration (nr) from the ISI distribution. We then fit the spike train with a PPM that predicts spike likelihood based on the time elapsed since the most recent of any spikes falling within the precedingnrmilliseconds. Finally, we compute the PSD of the model's residuals.Main results. We compared the residuals and shuffling methods' ability to enable accurate oscillation detection with flat baseline-assuming tests. Over synthetic data, the residuals method generally outperformed the shuffling method in classification of true- versus false-positive oscillatory power, principally due to enhanced sensitivity in sparse spike trains. In single-unit data from the internal globus pallidus (GPi) and ventrolateral anterior thalamus (VLa) of a parkinsonian monkey-in which alpha-beta oscillations (8-30 Hz) were anticipated-the residuals method reported the greatest incidence of significant alpha-beta power, with low firing rates predicting residuals-selective oscillation detection.Significance. These results encourage continued development of the residuals approach, to support more accurate oscillation detection. Improved identification of oscillations could promote improved disease models and therapeutic technologies.

Asunto(s)

RESUMEN

Irregular sampling of time series in electronic health records (EHRs) is one of the main challenges for developing machine learning models. Additionally, the pattern of missing values in certain clinical variables is not at random but depends on the decisions of clinicians and the state of the patient. Point process is a mathematical framework for analyzing event sequence data consistent with irregular sampling patterns. Our model, TEE4EHR, is a transformer event encoder (TEE) with point process loss that encodes the pattern of laboratory tests in EHRs. The utility of our TEE has been investigated in various benchmark event sequence datasets. Additionally, we conduct experiments on two real-world EHR databases to provide a more comprehensive evaluation of our model. Firstly, in a self-supervised learning approach, the TEE is jointly learned with an existing attention-based deep neural network, which gives superior performance in negative log-likelihood and future event prediction. Besides, we propose an algorithm for aggregating attention weights to reveal the events' interactions. Secondly, we transfer and freeze the learned TEE to the downstream task for the outcome prediction, where it outperforms state-of-the-art models for handling irregularly sampled time series. Furthermore, our results demonstrate that our approach can improve representation learning in EHRs and be useful for clinical prediction tasks.

Asunto(s)

RESUMEN

A factor constraining the elimination of dog-mediated human rabies is limited information on the size and spatial distribution of free-roaming dog populations (FRDPs). The aim of this study was to develop a statistical model to predict the size of free-roaming dog populations and the spatial distribution of free-roaming dogs in urban areas of Nepal, based on real-world dog census data from the Himalayan Animal Rescue Trust (HART) and Animal Nepal. Candidate explanatory variables included proximity to roads, building density, specific building types, human population density and normalised difference vegetation index (NDVI). A multivariable Poisson point process model was developed to estimate dog population size in four study locations in urban Nepal, with building density and distance from nearest retail food establishment or lodgings as explanatory variables. The proposed model accurately predicted, within a 95 % confidence interval, the surveyed FRDP size and spatial distribution for all four study locations. This model is proposed for further testing and refinement in other locations as a decision-support tool alongside observational dog population size estimates, to inform dog health and public health initiatives including rabies elimination efforts to support the 'zero by 30' global mission.

Asunto(s)

RESUMEN

The coexistence of multiple phytoplankton species despite their reliance on similar resources is often explained with mean-field models assuming mixed populations. In reality, observations of phytoplankton indicate spatial aggregation at all scales, including at the scale of a few individuals. Local spatial aggregation can hinder competitive exclusion since individuals then interact mostly with other individuals of their own species, rather than competitors from different species. To evaluate how microscale spatial aggregation might explain phytoplankton diversity maintenance, an individual-based, multispecies representation of cells in a hydrodynamic environment is required. We formulate a three-dimensional and multispecies individual-based model of phytoplankton population dynamics at the Kolmogorov scale. The model is studied through both simulations and the derivation of spatial moment equations, in connection with point process theory. The spatial moment equations show a good match between theory and simulations. We parameterized the model based on phytoplankters' ecological and physical characteristics, for both large and small phytoplankton. Defining a zone of potential interactions as the overlap between nutrient depletion volumes, we show that local species composition-within the range of possible interactions-depends on the size class of phytoplankton. In small phytoplankton, individuals remain in mostly monospecific clusters. Spatial structure therefore favours intra- over inter-specific interactions for small phytoplankton, contributing to coexistence. Large phytoplankton cell neighbourhoods appear more mixed. Although some small-scale self-organizing spatial structure remains and could influence coexistence mechanisms, other factors may need to be explored to explain diversity maintenance in large phytoplankton.

Asunto(s)

RESUMEN

The kppm is a standard procedure to estimate the parameters of the inhomogeneous Cox point process. However, the procedure cannot handle the problem when the models involve correlated covariates. In this study, we develop the kppmenet, the modified version of the kppm, for the inhomogeneous Cox point process involving correlated covariates by considering elastic net regularization. We compare the methodology in a simulation study and apply it to model major-shallow earthquake distribution in Sumatra, Indonesia. We conclude that the kppmenet outperforms kppm when correlated covariates are involved.

RESUMEN

Spatial heterogeneity of cells in liver biopsies can be used as biomarker for disease severity of patients. This heterogeneity can be quantified by non-parametric statistics of point pattern data, which make use of an aggregation of the point locations. The method and scale of aggregation are usually chosen ad hoc, despite values of the aforementioned statistics being heavily dependent on them. Moreover, in the context of measuring heterogeneity, increasing spatial resolution will not endlessly provide more accuracy. The question then becomes how changes in resolution influence heterogeneity indicators, and subsequently how they influence their predictive abilities. In this paper, cell level data of liver biopsy tissue taken from chronic Hepatitis B patients is used to analyze this issue. Firstly, Morisita-Horn indices, Shannon indices and Getis-Ord statistics were evaluated as heterogeneity indicators of different types of cells, using multiple resolutions. Secondly, the effect of resolution on the predictive performance of the indices in an ordinal regression model was investigated, as well as their importance in the model. A simulation study was subsequently performed to validate the aforementioned methods. In general, for specific heterogeneity indicators, a downward trend in predictive performance could be observed. While for local measures of heterogeneity a smaller grid-size is outperforming, global measures have a better performance with medium-sized grids. In addition, the use of both local and global measures of heterogeneity is recommended to improve the predictive performance.

Asunto(s)

RESUMEN

Point process data have become increasingly popular these days. For example, many of the data captured in electronic health records (EHR) are in the format of point process data. It is of great interest to study the association between a point process predictor and a scalar response using generalized functional linear regression models. Various generalized functional linear regression models have been developed under different settings in the past decades. However, existing methods can only deal with functional or longitudinal predictors, not point process predictors. In this article, we propose a novel generalized functional linear regression model for a point process predictor. Our proposed model is based on the joint modeling framework, where we adopt a log-Gaussian Cox process model for the point process predictor and a generalized linear regression model for the outcome. We also develop a new algorithm for fast model estimation based on the Gaussian variational approximation method. We conduct extensive simulation studies to evaluate the performance of our proposed method and compare it to competing methods. The performance of our proposed method is further demonstrated on an EHR dataset of patients admitted into the intensive care units of the Beth Israel Deaconess Medical Center between 2001 and 2008.

Asunto(s)

RESUMEN

A major challenge in the spatial analysis of multiplex imaging (MI) data is choosing how to measure cellular spatial interactions and how to relate them to patient outcomes. Existing methods to quantify cell-cell interactions do not scale to the rapidly evolving technical landscape, where both the number of unique cell types and the number of images in a dataset may be large. We propose a scalable analytical framework and accompanying R package, DIMPLE, to quantify, visualize, and model cell-cell interactions in the TME. By applying DIMPLE to publicly available MI data, we uncover statistically significant associations between image-level measures of cell-cell interactions and patient-level covariates.

RESUMEN

We consider a point process (PP) generated by superimposing an independent Poisson point process (PPP) on each line of a 2D Poisson line process (PLP). Termed PLP-PPP, this PP is suitable for modeling networks formed on an irregular collection of lines, such as vehicles on a network of roads and sensors deployed along trails in a forest. Inspired by vehicular networks in which vehicles connect with their nearest wireless base stations (BSs), we consider a random bipartite associator graph in which each point of the PLP-PPP is associated with the nearest point of an independent PPP through an edge. This graph is equivalent to the partitioning of PLP-PPP by a Poisson Voronoi tessellation (PVT) formed by an independent PPP. We first characterize the exact distribution of the number of points of PLP-PPP falling inside the ball centered at an arbitrary location in R2 as well as the typical point of PLP-PPP. Using these distributions, we derive cumulative distribution functions (CDFs) and probability density functions (PDFs) of kth contact distance (CD) and the nearest neighbor distance (NND) of PLP-PPP. As intermediate results, we present the empirical distribution of the perimeter and approximate distribution of the length of the typical chord of the zero-cell of this PVT. Using these results, we present two close approximations of the distribution of node degree of the random bipartite associator graph. In a vehicular network setting, this result characterizes the number of vehicles connected to each BS, which models its load. Since each BS has to distribute its limited resources across all the vehicles connected to it, a good statistical understanding of load is important for an efficient system design. Several applications of these new results to different wireless network settings are also discussed.

RESUMEN

Burst patterns, characterized by their temporal heterogeneity, have been observed across a wide range of domains, encompassing event sequences from neuronal firing to various facets of human activities. Recent research on predicting event sequences leveraged a Transformer based on the Hawkes process, incorporating a self-attention mechanism to capture long-term temporal dependencies. To effectively handle bursty temporal patterns, we propose a Burst and Memory-aware Transformer (BMT) model, designed to explicitly address temporal heterogeneity. The BMT model embeds the burstiness and memory coefficient into the self-attention module, enhancing the learning process with insights derived from the bursty patterns. Furthermore, we employed a novel loss function designed to optimize the burstiness and memory coefficient values, as well as their corresponding discretized one-hot vectors, both individually and jointly. Numerical experiments conducted on diverse synthetic and real-world datasets demonstrated the outstanding performance of the BMT model in terms of accurately predicting event times and intensity functions compared to existing models and control groups. In particular, the BMT model exhibits remarkable performance for temporally heterogeneous data, such as those with power-law inter-event time distributions. Our findings suggest that the incorporation of burst-related parameters assists the Transformer in comprehending heterogeneous event sequences, leading to an enhanced predictive performance.

RESUMEN

Understanding spatially varying survival is crucial for understanding the ecology and evolution of migratory animals, which may ultimately help to conserve such species. We develop an approach to estimate an annual survival probability function varying continuously in geographic space, if the recovery probability is constant over space. This estimate is based on a density function over continuous geographic space and the discrete age at death obtained from dead recovery data. From the same density function, we obtain an estimate for animal distribution in space corrected for survival, i.e., migratory connectivity. This is possible, when migratory connectivity can be separated from recovery probability. In this article, we present the method how spatially and continuously varying survival and the migratory connectivity corrected for survival can be obtained, if a constant recovery probability can be assumed reasonably. The model is a stepping stone in developing a model allowing for disentangling spatially heterogeneous survival and migratory connectivity corrected for survival from a spatially heterogeneous recovery probability. We implement the method using kernel density estimates in the R-package CONSURE. Any other density estimation technique can be used as an alternative. In a simulation study, the estimators are unbiased but show edge effects in survival and migratory connectivity. Applying the method to a real-world data set of European robins Erithacus rubecula results in biologically reasonable continuous heat-maps for survival and migratory connectivity.

RESUMEN

We present practical solutions to applying Gaussian-process (GP) methods to calculate spatial statistics for grid cells in large environments. GPs are a data efficient approach to inferring neural tuning as a function of time, space, and other variables. We discuss how to design appropriate kernels for grid cells, and show that a variational Bayesian approach to log-Gaussian Poisson models can be calculated quickly. This class of models has closed-form expressions for the evidence lower-bound, and can be estimated rapidly for certain parameterizations of the posterior covariance. We provide an implementation that operates in a low-rank spatial frequency subspace for further acceleration, and demonstrate these methods on experimental data.

Asunto(s)

RESUMEN

Recently, federated learning (FL) has been receiving great attention as an effective machine learning method to avoid the security issue in raw data collection, as well as to distribute the computing load to edge devices. However, even though wireless communication is an essential component for implementing FL in edge networks, there have been few works that analyze the effect of wireless networks on FL. In this paper, we investigate FL in small-cell networks where multiple base stations (BSs) and users are located according to a homogeneous Poisson point process (PPP) with different densities. We comprehensively analyze the effects of geographic node deployment on the model aggregation in FL on the basis of stochastic geometry-based analysis. We derive the closed-form expressions of coverage probability with tractable approximations and discuss the minimum required BS density for achieving a target model aggregation rate in small-cell networks. Our analysis and simulation results provide insightful information for understanding the behaviors of FL in small-cell networks; these can be exploited as a guideline for designing the network facilitating wireless FL.

RESUMEN

The organization and interaction between hepatocytes and other hepatic non-parenchymal cells plays a pivotal role in maintaining normal liver function and structure. Although spatial heterogeneity within the tumor micro-environment has been proven to be a fundamental feature in cancer progression, the role of liver tissue topology and micro-environmental factors in the context of liver damage in chronic infection has not been widely studied yet. We obtained images from 110 core needle biopsies from a cohort of chronic hepatitis B patients with different fibrosis stages according to METAVIR score. The tissue sections were immunofluorescently stained and imaged to determine the locations of CD45 positive immune cells and HBsAg-negative and HBsAg-positive hepatocytes within the tissue. We applied several descriptive techniques adopted from ecology, including Getis-Ord, the Shannon Index and the Morisita-Horn Index, to quantify the extent to which immune cells and different types of liver cells co-localize in the tissue biopsies. Additionally, we modeled the spatial distribution of the different cell types using a joint log-Gaussian Cox process and proposed several features to quantify spatial heterogeneity. We then related these measures to the patient fibrosis stage by using a linear discriminant analysis approach. Our analysis revealed that the co-localization of HBsAg-negative hepatocytes with immune cells and the co-localization of HBsAg-positive hepatocytes with immune cells are equally important factors for explaining the METAVIR score in chronic hepatitis B patients. Moreover, we found that if we allow for an error of 1 on the METAVIR score, we are able to reach an accuracy of around 80%. With this study we demonstrate how methods adopted from ecology and applied to the liver tissue micro-environment can be used to quantify heterogeneity and how these approaches can be valuable in biomarker analyses for liver topology.

Asunto(s)

RESUMEN

The tumor microenvironment (TME) is a complex ecosystem containing tumor cells, other surrounding cells, blood vessels, and extracellular matrix. Recent advances in multiplexed imaging technologies allow researchers to map several cellular markers in the TME at the single cell level while preserving their spatial locations. Evidence is mounting that cellular interactions in the TME can promote or inhibit tumor development and contribute to drug resistance. Current statistical approaches to quantify cell-cell interactions do not readily scale to the outputs of new imaging technologies which can distinguish many unique cell phenotypes in one image. We propose a scalable analytical framework and accompanying R package, DIMPLE, to quantify, visualize, and model cell-cell interactions in the TME. In application of DIMPLE to publicly available MI data, we uncover statistically significant associations between image-level measures of cell-cell interactions and patient-level covariates.

RESUMEN

We revisit a spatial metapopulation model on continuous space as a stochastic point pattern dynamics. In the model, local patches as points are distributed with a certain spatial configuration and status of each patch changes stochastically between occupied and empty: an occupied patch becomes empty by local extinction and an empty patch becomes occupied both by local and global colonization. We carry out simulation analysis and derive an analytical model in terms of singlet, pair and triplet probabilities that describe the stochastic dynamics. Using a simple closure that approximates triplet probabilities by singlet and pair probabilities, we show that equilibrium singlet and pair probabilities can be analytically derived. The derived equilibrium properties successfully describe simulation results under a certain condition where the range of local colonization and the proportion of global colonization play key roles. Our model is an extension of the classical non-spatial Levins model to a spatially explicit metapopulation model. We appeal the advantage of point pattern approach to study spatial dynamics in general ecology and call for the need to deepen our understanding of mathematical tools to explore point pattern dynamics.

RESUMEN

Swarm density plays a key role in the performance of a robot swarm, which can be averagely measured by swarm size and the area of a workspace. In some scenarios, the swarm workspace may not be fully or partially observable, or the swarm size may decrease over time due to out-of-battery or faulty individuals during operation. This can result in the average swarm density over the whole workspace being unable to be measured or changed in real-time. The swarm performance may not be optimal due to unknown swarm density. If the swarm density is too low, inter-robot communication will rarely be established, and robot swarm cooperation will not be effective. Meanwhile, a densely-packed swarm compels robots to permanently solve collision avoidance issues rather than performing the main task. To address this issue, in this work, the distributed algorithm for collective cognition on the average global density is proposed. The main idea of the proposed algorithm is to help the swarm make a collective decision on whether the current global density is larger, smaller or approximately equal to the desired density. During the estimation process, the swarm size adjustment is acceptable for the proposed method in order to reach the desired swarm density.