Title: Nonparametric Statistics in Omics Applications

Organizers: Ahmed A. Metwally (Stanford University); Alan Perez-Rathke (University of Illinois at Chicago)

Overview: With the rapidly increasing size and complexity of omics data sets, nonparametric methods are emerging as principal tools for biomedical analysis. Nonparametric statistics have the advantages of making minimal distributional assumptions and can scale to fit the complexity of the data. The purpose of this session is to highlight novel applications of nonparametric methods, such as nonparametric longitudinal analysis, bootstrapping, approximate Bayesian inference, and Dirichlet processes, towards omics and biomedical data.


Title: AI techniques for multi-modality medical big data

Organizers: Ye Li (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China)

Overview: Artificial intelligence has been widely used in computer-aided diagnosis based on medical images in recent years, and also been applied for EMR big data analysis. With the development of wearable devices, which can monitor physiological signals continuously out of the hospitals, the fusion of physiological signals, EMRs and medical imaging will provide novel insights for medical data analysis. Compared with single-mode data mining, the method based on multi-modality medical data fusion can effectively mine the relationship between different medical data and improve the effect of assistant diagnosis. This special session will focus on AI techniques for multi-modality medical big data analysis, including but not limited to: data mining of EMRs, information fusion of EMRs, medical images, physiological data, and genomic data for diagnosis and early-warning of diseases.


Title: Genome Security & Privacy

Organizers: Gamze Gursoy (Yale University); Haixu Tang (Indiana University)

Overview: Data privacy is an important topic of debate crossing many different fields such as ethics, sociology, law, political science and forensic science. With decreasing cost of DNA sequencing technologies, the number and the size of the available genomic data have exponentially increased and become available to wider audiences. Hence, privacy of individuals’ genomic data has recently emerged as one of the major foci of studies on privacy as availability of genetic information gives rise to privacy concerns such that genetic predisposition to diseases may bias insurance companies or create unlawful discrimination by employers. Recently it has been also showed that not only DNA sequencing of individuals but also high throughput molecular phenotype datasets such as functional genomic and metabolomics measurements or even microbiome measurements, increased the number of quasi-identifiers for participating individuals can be used by adversaries for reidentification purposes. These results indicate that privacy concerns over sharing personalized biological data will increase quickly with the increase in the number of genetic and ancestry testing companies, which collect and distribute very large amount of health related data. These include genetic information (such as 23andme) or health and fitness tracking information (such as fitbit). The data collection and sharing methods that these companies use call for a public discussion of privacy considerations around these new concepts.


Title: Wearable Sensor Informatics for Cardiopulmonary Monitoring

Organizers: Inan Omer (Georgia Institute of Technology); Jin-Oh Hahn (University of Maryland)

Overview: While wearable sensing for cardiopulmonary health is typically covered in standard sessions at BHI, there is a compelling need to bring the community together for a special session aimed at (1) elucidating some of recent scientific discoveries in the field; (2) familiarizing the community with new sensing technologies, materials, and analytics methodologies; and (3) understanding some of the emerging sensing modalities and signal processing/system analysis algorithms. We anticipate that bringing these speakers together to discuss the latest trends in their respective areas will lead to synergistic opportunities for collaboration, and will also result in productive and exciting conversations at BHI for future research opportunities.


Title: Internet of Things and Machine Learning for Health Informatics

Organizers: Wei Chen (Director of Centre for Intelligent Medical Electronics (CIME), Department of Electronic Engineering, School of Information Science and Technology, Fudan University); Benny Lo (Senior Lecturer – Hamlyn Centre, the Department of Surgery and Cancer Imperial College London)

Overview: <any challenges exit in health monitoring and management, such as continuous, accurate, and comfortable monitoring of multi-parameters, early detection and warning of diseases, as well as the interaction with environments. The challenges in healthcare raise health risks and imposes significant economic and social burden. Thus, seeking for the innovative solutions and new technologies to address these issues is very important. The development of modern sensors, Internet of Things, advanced materials, machine learning and data fusion technology has inspired the innovation on intelligent designs for healthcare.


Title: Decision-support computing by data-driven and AI-based approaches for healthcare

Organizers: Guillaume Bouleux (Univ Lyon, INSA-Lyon, DISP EA); Sondes Chaabane (University of Valenciennes, LAMIH, UPHF, Le Mont Houy)

Overview: The adoption of artificial intelligence (AI) in healthcare is on the rise and solving a variety of problems for patients, hospitals and healthcare industry overall. In this context, many sources of health data can be used, such as electronic health records (EHRs) or medico-administrative data. Recent studies have shown that secondary use of EHRs has enabled data-driven detection and prediction of abnormal paients flow at Emergency Department and some other works have yielded to the identification of certain pathologies through clustering approaches. Due to huge records at the Hospital, the prediction of patient pathway can also be tackled by deep learnig methods. All these methods concern the public health and try to imporove patient care by giving to pactitionners a web application or software to help in decision making.