A modified SEIR model to predict the COVID-19 outbreak in Spain and Italy: simulating control scenarios and multi-scale epidemics

AbstractAfter the spread of SARS-CoV-2 epidemic out of China, evolution in the pandemic worldwide shows dramatic differences among countries. In Europe, the situation of Italy first and later Spain has generated great concern, and despite other countries show better prospects, large uncertainties yet remain on the future evolution and the efficacy of containment, mitigation or attack strategies. Here we applied a modified SEIR compartmental model accounting for the spread of infection during the latent period, in which we also incorporate effects of varying proportions of containment. We fit data to quarantined populations in order to account for the uncertainties in case reporting and study the scenario projections for the 17 individual regions (CCAA). Results indicate that with data for March 23, the epidemics follows an evolution similar to the isolation of 1, 5 percent of the population and if there were no effects of intervention actions it might reach a maximum over 1.4M infected around April27. The effect on the epidemics of the ongoing partial confinement measures is yet unknown (an update of results with data until March 31st is included), but increasing the isolation around ten times more could drastically reduce the peak to over 100k cases by early April, while each day of delay in taking this hard containment scenario represents an 90 percent increase of the infected population at the peak. Dynamics at the sub aggregated levels of CCAA show epidemics at the different levels of progression with the most worrying situation in Madrid an Catalonia. Increasing alpha values up to 10 times, in addition to a drastic reduction in clinical cases, would also more than halve the number of deaths. Updates for March 31st simulations indicate a substantial reduction in burden is underway. A similar approach conducted for Italy pre- and post-interventions also begins to suggest substantial reduction in both infected and deaths has been achieved, showing the efficacy of drastic social distancing interventions.

Download Full-text

Geographic Disparities in Health

Health in Japan ◽

10.1093/oso/9780198848134.003.0017 ◽

2020 ◽

pp. 265-280

Author(s):

Tomoki Nakaya ◽

Tomoya Hanibuchi

Keyword(s):

Social Inequalities ◽

Residential Areas ◽

Metropolitan Regions ◽

Multi Scale ◽

The World ◽

Socioeconomic Segregation ◽

Supportive Environments ◽

Post War ◽

Neighbourhood Scale ◽

Different Levels

This chapter highlights the geographical aspects of health disparities in Japan at different levels, from the 47 prefectures nationally to the neighbourhood level. In the post-war period, Japan has successfully attained the longest life expectancy in the world. At the same time, it has substantially reduced geographical disparities among the prefectures. This indicates that reducing such disparities in living standards may also be related to improving the health of a country’s entire population. However, disparities of health have appeared among populations living in socially segmented areas in large neighbourhoods of metropolitan regions. Such neighbourhood-scale disparities in health are associated with a number of environmental characteristics of Japanese neighbourhoods reflecting socioeconomic segregation and development histories of residential areas. In the era of a super-aging society that contains the threat of re-widening social inequalities, Japan faces challenges to build health-supportive environments for tackling multi-scale disparities.

Download Full-text

Adversarial Training for Community Question Answer Selection Based on Multi-Scale Matching

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v33i01.3301395 ◽

2019 ◽

Vol 33 ◽

pp. 395-402

Author(s):

Xiao Yang ◽

Madian Khabsa ◽

Miaosen Wang ◽

Wei Wang ◽

Ahmed Hassan Awadallah ◽

...

Keyword(s):

Question Answering ◽

State Of The Art ◽

Classification Problem ◽

Classification Model ◽

Training Set ◽

Community Based ◽

Multi Scale ◽

Adversarial Training ◽

Source Of Information ◽

Different Levels

Community-based question answering (CQA) websites represent an important source of information. As a result, the problem of matching the most valuable answers to their corresponding questions has become an increasingly popular research topic. We frame this task as a binary (relevant/irrelevant) classification problem, and present an adversarial training framework to alleviate label imbalance issue. We employ a generative model to iteratively sample a subset of challenging negative samples to fool our classification model. Both models are alternatively optimized using REINFORCE algorithm. The proposed method is completely different from previous ones, where negative samples in training set are directly used or uniformly down-sampled. Further, we propose using Multi-scale Matching which explicitly inspects the correlation between words and ngrams of different levels of granularity. We evaluate the proposed method on SemEval 2016 and SemEval 2017 datasets and achieves state-of-the-art or similar performance.

Download Full-text

Hi-Fi: Hierarchical Feature Integration for Skeleton Detection

Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2018/166 ◽

2018 ◽

Cited By ~ 13

Author(s):

Kai Zhao ◽

Wei Shen ◽

Shanghua Gao ◽

Dandan Li ◽

Ming-Ming Cheng

Keyword(s):

Performance Improvement ◽

State Of The Art ◽

Natural Images ◽

Feature Integration ◽

Detection Problem ◽

Multi Scale ◽

Integration Mechanism ◽

Object Parts ◽

High Level ◽

Different Levels

In natural images, the scales (thickness) of object skeletons may dramatically vary among objects and object parts. Thus, robust skeleton detection requires powerful multi-scale feature integration ability. To address this issue, we present a new convolutional neural network (CNN) architecture by introducing a novel hierarchical feature integration mechanism, named Hi-Fi, to address the object skeleton detection problem. The proposed CNN-based approach intrinsically captures high-level semantics from deeper layers, as well as low-level details from shallower layers. By hierarchically integrating different CNN feature levels with bidirectional guidance, our approach (1) enables mutual refinement across features of different levels, and (2) possesses the strong ability to capture both rich object context and high-resolution details. Experimental results show that our method significantly outperforms the state-of-the-art methods in terms of effectively fusing features from very different scales, as evidenced by a considerable performance improvement on several benchmarks.

Download Full-text

Describing the COVID-19 Outbreak Fitting Modified SIR models to Data

10.1101/2020.04.29.20084285 ◽

2020 ◽

Author(s):

Aldo Ianni ◽

Nicola Rossi

Keyword(s):

Time Evolution ◽

Compartmental Model ◽

Social Distancing ◽

Future Evolution ◽

The Social ◽

Model Predictions ◽

Official Sources ◽

Far Future

AbstractIn this paper we fit simple modifications of the SIR compartmental model to the COVID-19 outbreak data, available from official sources for Italy and other countries. Even if the complexity of the pandemic can not be easily modelled, we show that our model, at present, describes the time evolution of the data in spite of the application of the social distancing and lock-down procedure. Finally, we discuss the reliability of the model predictions, under certain conditions, for estimating the near and far future evolution of the COVID-19 outbreak. The conditions for the applicability of the proposed models are discussed.

Download Full-text

Efficient, high-performance semantic segmentation using multi-scale feature extraction

PLoS ONE ◽

10.1371/journal.pone.0255397 ◽

2021 ◽

Vol 16 (8) ◽

pp. e0255397

Author(s):

Moritz Knolle ◽

Georgios Kaissis ◽

Friederike Jungmann ◽

Sebastian Ziegelmayer ◽

Daniel Sasse ◽

...

Keyword(s):

Deep Learning ◽

Graphics Processing Units ◽

Substantial Reduction ◽

Image Features ◽

Tumor Segmentation ◽

Processing Unit ◽

Central Processing ◽

Multi Scale ◽

Computational Performance ◽

Wide Range

The success of deep learning in recent years has arguably been driven by the availability of large datasets for training powerful predictive algorithms. In medical applications however, the sensitive nature of the data limits the collection and exchange of large-scale datasets. Privacy-preserving and collaborative learning systems can enable the successful application of machine learning in medicine. However, collaborative protocols such as federated learning require the frequent transfer of parameter updates over a network. To enable the deployment of such protocols to a wide range of systems with varying computational performance, efficient deep learning architectures for resource-constrained environments are required. Here we present MoNet, a small, highly optimized neural-network-based segmentation algorithm leveraging efficient multi-scale image features. MoNet is a shallow, U-Net-like architecture based on repeated, dilated convolutions with decreasing dilation rates. We apply and test our architecture on the challenging clinical tasks of pancreatic segmentation in computed tomography (CT) images as well as brain tumor segmentation in magnetic resonance imaging (MRI) data. We assess our model’s segmentation performance and demonstrate that it provides performance on par with compared architectures while providing superior out-of-sample generalization performance, outperforming larger architectures on an independent validation set, while utilizing significantly fewer parameters. We furthermore confirm the suitability of our architecture for federated learning applications by demonstrating a substantial reduction in serialized model storage requirement as a surrogate for network data transfer. Finally, we evaluate MoNet’s inference latency on the central processing unit (CPU) to determine its utility in environments without access to graphics processing units. Our implementation is publicly available as free and open-source software.

Download Full-text

Neurobehavioral Differences across Strategies Used in Cardiac Recognition

10.31234/osf.io/649p7 ◽

2019 ◽

Author(s):

Lilla Hodossy ◽

Manos Tsakiris

Keyword(s):

Active Control ◽

Functional Approach ◽

The State ◽

Conscious Perception ◽

Cortical Processing ◽

Future Evolution ◽

Passive Strategies ◽

Different Levels

The maintenance of psycho-physiological stability requires the ability to infer the state of our body (interoception) and to predict its future evolution. Yet standard measures of interoception do not have this functional approach as they are typically limited to the conscious perception of single heartbeats. We here present a new biofeedback paradigm to explore the effect of three strategies (i.e. exteroceptive, active or passive interoceptive) on interoceptive inference – defined here as the ability to recognize one’s own heart. We observed an increase of cardiac recognition and a more pronounced cortical processing of heartbeats across both interoceptive strategies as compared to the exteroceptive one. We also observed the highest level of metacognition at the active, control-based interoceptive strategy. Strategy-specific cardiac recognition was linked to the modulation of cortical processing of heartbeats, exclusively in the passive interoceptive condition. We suggest that while both active and passive strategies increase the precision of the interoceptive channel, they exert distinct influences on different levels of the interoceptive hierarchy.

Download Full-text

A Study on the Effects of Containment Policies and Vaccination on the Spread of SARS-CoV-2

10.1101/2020.10.16.20213835 ◽

2020 ◽

Author(s):

Vahid S. Bokharaie

Keyword(s):

Age Structure ◽

Compartmental Model ◽

Herd Immunity ◽

Age Groups ◽

Age Group ◽

Model Output ◽

Monotone Systems ◽

Different Populations ◽

Different Levels

AbstractThis paper presents a method to predict the spread of the SARS-CoV-2 in a population with a known age-structure, and then, to quantify the effects of various containment policies, including those policies that affect each age-group differently. The model itself is a compartmental model in which each compartment is divided into a number of age-groups. The parameter of the model are estimated using an optimisation scheme and some known results from the theory of monotone systems such that the model output agrees with some collected data on the spread of SARS-CoV-2.To highlight the strengths of this framework, a few case studies are presented in which different populations are subjected to different containment strategies. They include cases in which the containment policies switch between scenarios with different levels of severity. Then a case study on herd immunity due to vaccination is presented. And then it is shown how we can use this framework to optimality distribute a limited number of vaccine units in a given population to maximise their impact and lower the total number of infectious individuals.MSC subclass92C60, 92C50

Download Full-text

Multi-scale verification of distributed synchronisation

Formal Methods in System Design ◽

10.1007/s10703-020-00347-z ◽

2020 ◽

Vol 55 (3) ◽

pp. 171-221

Author(s):

Paul Gainer ◽

Sven Linker ◽

Clare Dixon ◽

Ullrich Hustadt ◽

Michael Fisher

Keyword(s):

Distributed Systems ◽

Coupled Oscillators ◽

Population Based ◽

Levels Of Abstraction ◽

Biologically Inspired ◽

Collective Models ◽

Multi Scale ◽

Pulse Coupled Oscillators ◽

The Individual ◽

Different Levels

AbstractAlgorithms for the synchronisation of clocks across networks are both common and important within distributed systems. We here address not only the formal modelling of these algorithms, but also the formal verification of their behaviour. Of particular importance is the strong link between the very different levels of abstraction at which the algorithms may be verified. Our contribution is primarily the formalisation of this connection between individual models and population-based models, and the subsequent verification that is then possible. While the technique is applicable across a range of synchronisation algorithms, we particularly focus on the synchronisation of (biologically-inspired) pulse-coupled oscillators, a widely used approach in practical distributed systems. For this application domain, different levels of abstraction are crucial: models based on the behaviour of an individual process are able to capture the details of distinguished nodes in possibly heterogenous networks, where each node may exhibit different behaviour. On the other hand, collective models assume homogeneous sets of processes, and allow the behaviour of the network to be analysed at the global level. System-wide parameters may be easily adjusted, for example environmental factors inhibiting the reliability of the shared communication medium. This work provides a formal bridge across the “abstraction gap” separating the individual models and the population-based models for this important class of synchronisation algorithms.

Download Full-text

PSANet: Pyramid Splitting and Aggregation Network for 3D Object Detection in Point Cloud

Sensors ◽

10.3390/s21010136 ◽

2020 ◽

Vol 21 (1) ◽

pp. 136

Author(s):

Fangyu Li ◽

Weizheng Jin ◽

Cien Fan ◽

Lian Zou ◽

Qingsheng Chen ◽

...

Keyword(s):

Object Detection ◽

Point Clouds ◽

Autonomous Driving ◽

Feature Maps ◽

3D Object ◽

Multi Scale ◽

Backbone Network ◽

3D Object Detection ◽

Different Levels ◽

Fine Branch

3D object detection in LiDAR point clouds has been extensively used in autonomous driving, intelligent robotics, and augmented reality. Although the one-stage 3D detector has satisfactory training and inference speed, there are still some performance problems due to insufficient utilization of bird’s eye view (BEV) information. In this paper, a new backbone network is proposed to complete the cross-layer fusion of multi-scale BEV feature maps, which makes full use of various information for detection. Specifically, our proposed backbone network can be divided into a coarse branch and a fine branch. In the coarse branch, we use the pyramidal feature hierarchy (PFH) to generate multi-scale BEV feature maps, which retain the advantages of different levels and serves as the input of the fine branch. In the fine branch, our proposed pyramid splitting and aggregation (PSA) module deeply integrates different levels of multi-scale feature maps, thereby improving the expressive ability of the final features. Extensive experiments on the challenging KITTI-3D benchmark show that our method has better performance in both 3D and BEV object detection compared with some previous state-of-the-art methods. Experimental results with average precision (AP) prove the effectiveness of our network.

Download Full-text

MULTI-SCALE ISO-SURFACE EXTRACTION FOR VOLUME VISUALIZATION

International Journal of Image and Graphics ◽

10.1142/s0219467806002185 ◽

2006 ◽

Vol 06 (02) ◽

pp. 173-185 ◽

Cited By ~ 3

Author(s):

SHIAOFEN FANG ◽

MARWAN ADADA

Keyword(s):

Large Volume ◽

Volume Visualization ◽

Scale Space ◽

Histogram Analysis ◽

Zero Crossing ◽

Surface Extraction ◽

Multi Scale ◽

Interactive Environment ◽

Different Levels ◽

Boundary Surfaces

This paper describes a new multi-scale approach for the extraction of iso-surfaces from volume datasets. The goal is to automatically identify iso-surfaces that best approximate the boundary surfaces at different levels of details. Using histogram analysis, iso-values are extracted from histograms of boundary voxels defined by gradient thresholding or zero-crossing boundaries. Multi-scale smoothing of the histogram using Gaussian filters of various sizes allows the iso-surfaces to be defined hierarchically over a scale space map. It provides an interactive environment and volume navigation tools for the exploration of large volume datasets by visualizing the layers of the volume space in a multi-scale manner.

Download Full-text