A5081476701- Human Mobility and Location-Based Analysis
- Data-Driven Disease Surveillance
- COVID-19 epidemiological studies
- Impact of Light on Environment and Health
- Urban Transport and Accessibility
- Complex Network Analysis Techniques
- SARS-CoV-2 and COVID-19 Research
- Mobile Crowdsensing and Crowdsourcing
- Anomaly Detection Techniques and Applications
- Evolution and Genetic Dynamics
- Food Safety and Hygiene
- Privacy-Preserving Technologies in Data
- Time Series Analysis and Forecasting
- Food Supply Chain Traceability
- Ethics in Clinical Research
- Data Management and Algorithms
- Influenza Virus Research Studies
- Machine Learning in Healthcare
- Land Use and Ecosystem Services
- Viral Infections and Outbreaks Research
- Context-Aware Activity Recognition Systems
- Transportation Planning and Optimization
- Tactile and Sensory Interactions
- Urban, Neighborhood, and Segregation Studies
- Data Stream Mining Techniques
Google (United States)
2013-2022
University of Rochester
2010-2021
Location plays an essential role in our lives, bridging online and offline worlds. This paper explores the interplay between people's location, interactions, their social ties within a large real-world dataset. We present evaluate Flap, system that solves two intimately related tasks: link location prediction networks. For prediction, Flap infers by considering patterns friendship formation, content of messages, user location. show while each component is weak predictor alone, combining them...
Better relaxing lockdown together Even during a pandemic, all countries—even islands—are dependent in one way or another on their neighbors. Without coordinated relaxation of nonpharmaceutical interventions (NPIs) among the most closely connected countries, it is difficult to envisage maintaining control infectious viruses such as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Ruktanonchai et al. used mobility data from smartphones estimate movements between administrative...
Research in computational epidemiology to date has concentrated on coarse-grained statistical analysis of populations, often synthetic ones. By contrast, this paper focuses fine-grained modeling the spread infectious diseases throughout a large real-world social network. Specifically, we study roles that ties and interactions between specific individuals play progress contagion. We focus public Twitter data, where find for every health-related message there are more than 1,000 unrelated This...
The recent trend of rapid urbanization makes it imperative to understand urban characteristics such as infrastructure, population distribution, jobs, and services that play a key role in livability sustainability. A healthy debate exists on what constitutes optimal structure regarding cities, interpolating, for instance, between mono- poly-centric organization. Here anonymous aggregated flows generated from three hundred million users, opted-in Location History, are used extract global...
Researchers have begun to mine social network data in order predict a variety of social, economic, and health related phenomena. While previous work has focused on predicting aggregate properties, such as the prevalence seasonal influenza given country, we consider task fine-grained prediction specific people from noisy incomplete data. We construct probabilistic model that can if when an individual will fall ill with high precision good recall basis his ties co-locations other people,...
Abstract Privacy protection is paramount in conducting health research. However, studies often rely on data stored a centralized repository, where analysis done with full access to the sensitive underlying content. Recent advances federated learning enable building complex machine-learned models that are trained distributed fashion. These techniques facilitate calculation of research study endpoints such private never leaves given device or healthcare system. We show—on diverse set single...
Much work has been done on predicting where is one going to be in the immediate future, typically within next hour. By contrast, we address open problem of human mobility far into a scale months and years. We propose an efficient nonparametric method that extracts significant robust patterns location data, learns their associations with contextual features (such as day week), subsequently leverages this information predict most likely at any given time future. The entire process formulated...
The SARS-CoV-2 Delta (Pango lineage B.1.617.2) variant of concern spread globally, causing resurgences COVID-19 worldwide1,2. emergence the in UK occurred on background a heterogeneous landscape immunity and relaxation non-pharmaceutical interventions. Here we analyse 52,992 genomes from England together with 93,649 rest world to reconstruct quantify its introduction regional dissemination across context changing travel social restrictions. Using analysis human movement, contact tracing...
Real-time captioning provides deaf and hard of hearing people immediate access to spoken language enables participation in dialogue with others. Low latency is critical because it allows speech be paired relevant visual cues. Currently, the only reliable source real-time captions are expensive stenographers who must recruited advance trained use specialized keyboards. Automatic recognition (ASR) less available on-demand, but its low accuracy, high noise sensitivity, need for training...
Research on human computation and crowdsourcing has concentrated tasks that can be accomplished remotely over the Internet. We introduce a general class of problems we call crowdphysics (CP)---crowdsourcing require people to collaborate synchronize both in time physical space. As an illustrative example, focus crowd-powered delivery service---a specific CP instance where go about their daily lives, but have opportunity carry packages delivered locations or individuals. Each package is handed...
Given the rapid recent trend of urbanization, a better understanding how urban infrastructure mediates socioeconomic interactions and economic systems is vital importance. While accessibility location-enabled devices as well large-scale datasets human activities, has fueled significant advances in our understanding, there little agreement on linkage between status its influence movement patterns, particular, role inequality. Here, we analyze heavily aggregated anonymized summary global...
Abstract Human mobility is a primary driver of infectious disease spread. However, existing data limited in availability, coverage, granularity, and timeliness. Data-driven forecasts dynamics are crucial for decision-making by health officials private citizens alike. In this work, we focus on machine-learned anonymized map (hereon referred to as AMM) aggregated over hundreds millions smartphones evaluate its utility forecasting epidemics. We factor AMM into metapopulation model...
Infectious disease forecasting has been a key focus in the recent past owing to COVID-19 pandemic and proved be an important tool controlling pandemic. With advent of reliable spatiotemporal data, graph neural network models have able successfully model inter-relation between cross-region signals produce quality forecasts, but like most deep-learning they do not explicitly incorporate underlying causal mechanisms. In this work, we employ mechanistic guide learning embeddings propose novel...
Abstract The ongoing SARS-CoV-2 pandemic has been holding the world hostage for several years now. Mobility is key to viral spreading and its restriction main non-pharmaceutical interventions fight virus expansion. Previous works have shown a connection between structural organization of cities movement patterns their residents. This puts urban centers in focus epidemic surveillance interventions. Here we show that flows tremendous impact on disease amenability different mitigation...
Machine learning has become an increasingly powerful tool for solving complex problems, and its application in public health been underutilized. The objective of this study is to test the efficacy a machine-learned model foodborne illness detection real-world setting. To end, we built FINDER, real-time using anonymous aggregated web search location data. We computed fraction people who visited particular restaurant later searched terms indicative food poisoning identify potentially unsafe...
Foodborne illness afflicts 48 million people annually in the US alone. More than 128,000 are hospitalized and 3000 die from infection. While preventable with proper food safety practices, traditional restaurant inspection process has limited impact given predictability low frequency of inspections, dynamic nature kitchen environment. Despite this reality, remained largely unchanged for decades. CDC even identified as one seven “winnable battles”; however, progress to date been limited. In...
Recent research has shown that surprisingly rich models of human behavior can be learned from GPS (positional) data. However, most to date concentrated on modeling single individuals or aggregate statistical properties groups people. Given noisy real-world data, we---in contrast---consider the problem and recognizing activities involve multiple related playing a variety roles. Our test domain is game capture flag---an outdoor involves many distinct cooperative competitive joint activities....
Research in computational epidemiology to date has concentrated on estimating summary statistics of populations and simulated scenarios disease outbreaks. Detailed studies have been limited small domains, as scaling the methods involved poses considerable challenges. By contrast, we model associations a large collection social environmental factors with health particular individuals. Instead relying surveys, apply scalable machine learning techniques noisy data mined from online media infer...
Computational approaches to health monitoring and epidemiology continue evolve rapidly. We present an end-to-end system, nEmesis, that automatically identifies restaurants posing public risks. Leveraging a language model of Twitter users' online communication, nEmesis finds individuals who are likely suffering from foodborne illness. People's visits modeled by matching GPS data embedded in the messages with restaurant addresses. As result, we can assign each venue "health score" based on...
Recent research has shown that surprisingly rich models of human activity can be learned from GPS (positional) data. However, most effort to date concentrated on modeling single individuals or statistical properties groups people. Moreover, prior work focused solely actual successful executions (and not failed attempted executions) the activities interest. We, in contrast, take task understanding interactions, and intentions noisy sensor data a fully relational multi-agent setting. We use...
Foodborne illness afflicts 48 million people annually in the U.S. alone. Over 128,000 are hospitalized and 3,000 die from infection. While preventable with proper food safety practices, traditional restaurant inspection process has limited impact given predictability low frequency of inspections, dynamic nature kitchen environment. Despite this reality, remained largely unchanged for decades. We apply machine learning to Twitter data develop a system that automatically detects venues likely...
Disease dynamics, human mobility, and public policies co-evolve during a pandemic such as COVID-19. Understanding dynamic mobility changes spatial interaction patterns are crucial for understanding forecasting COVID-19 dynamics. We introduce novel graph-based neural network(GNN) to incorporate global aggregated flows better of the impact on dynamics well disease propose recurrent message passing graph network that embeds spatio-temporal daily state-level new confirmed cases forecasting. This...