Distinguished Seminars

Jan. 8, 2018

Edward Chang of HTC

“Advancing Healthcare with AI and VR”
Edward Chang
President of Research & Healthcare (DeepQ) at HTC
Location and time: TBD

Quality, cost, and accessibility form an iron triangle that has prevented healthcare from achieving accelerated advancement in the last few decades. Improving any one of the three metrics may lead to degradation of the other two. However, thanks to recent breakthroughs in artificial intelligence (AI) and virtual reality (VR), this iron triangle can finally be shattered. In this talk, I will share the experience of developing DeepQ, an AI platform for AI-assisted diagnosis and VR-facilitated surgery. I will present three healthcare initiatives we have undertaken since 2012: Healthbox, Tricorder, and VR surgery, and explain how AI and VR play pivotal roles in improving diagnosis accuracy and treatment effectiveness. And more specifically, how we have dealt with not only big data analytics, but also small data learning, which is typical in the medical domain. The talk concludes with roadmaps and a list of open research issues in signal processing and AI to achieve precision medicine and surgery.

Dec. 11, 2017

Mark MapstoneMassimo Fiandaca

“Biomarker Development for Alzheimer’s Disease: Opportunities, Challenges, and the Need for Big Data Analytics”
Massimo Fiandaca and Mark Mapstone
Department of Neurology, UCI School of Medicine
4 p.m. in Donald Bren Hall 4011

Alzheimer’s disease (AD) is a devastating brain illness that slowly robs older adults of their memory and independence. We currently have no cures or disease modifying therapies and only temporary and minimally effective treatments for certain symptoms. Biomarkers obtained from blood or other body fluids may provide key insights into the underlying pathobiology of AD, before the illness becomes evident. Biomarker analysis of the preclinical phase may inform us regarding the relevant pathobiology and thereby suggest novel therapies that might mitigate or prevent AD.

One of the major hurdles to biomarker research is the complexity of the clinical pathophysiology and by extension, the density of the data. It is clear that multiple levels of clinical, lifestyle, environmental, and biological information, over time, must be considered for accurate and early detection of preclinical (asymptomatic) AD. A multimodal “big data” analytic approach will be required to manage these mutually informative, but distinct datasets, in developing a better understanding of AD, and other human conditions.

In this talk, Drs. Fiandaca and Mapstone will describe their work in preclinical AD biomarker development. They will give an overview of their recent findings related to metabolomic markers of AD and present the potential for integration of other relevant –omics, including proteomics, transcriptomics, genomics, and epigenomics for enhanced understanding of the underlying pathobiology. They will also present work on metabolomics of successful cognitive aging which will be important when considering approaches to primary prevention of AD. This talk hopes to present the unique opportunities and challenges associated with complex multidimensional datasets and the growing need for big data analytic approaches in medicine.

Oct. 30, 2017

Ophir Frieder

“Computing Medicine”
Ophir Frieder
Professor of Computer Science; Professor of Biostatistics, Bioinformatics and Biomathematics, Georgetown University
4 p.m. in Donald Bren Hall 4011

Computing continues to change the landscape of nearly all domains, medicine included. For instance, drug resistance is predicted and avoided via data mining applications; radiological reading errors are detected and prevented via natural language processing; and disease outbreak is detected early via text mining techniques. These are but just some examples where computing is reshaping medical practice. Specifically, we describe the monitoring of social media to detect disease outbreak and describe the implications of such surveillance schemes to healthcare planning for a major children-focused hospital. We continue by describing how conventional mining approaches significantly improve urinary tract infection treatment plans as developed jointly with and for another major hospital. Finally, we describe automated means for the detection of differences in radiological readings and describe how such detection schemes are used in yet a third major hospital.

Oct. 16, 2017

Aidong Zhang

“Data Driven Self-Learning for Knowledge Discovery In Health”
Aidong Zhang
Distinguished Professor of Computer Science and Engineering, State University of New York
4 p.m. in Donald Bren Hall 4011

With the growth of world wide web and large-scale digitization of documents, we are overwhelmed with massive information, formally through publication of various scientific journals or informally through internet. As an example, consider MEDLINE, a premier bibliographic database in life sciences, with currently more than 23 million references from approximately 5,600 worldwide journals. As a consequence, Literature Based Discovery (LBD) has become a sub-field of Text Mining that leverages these published articles to formulate hypotheses. This area of research, which incorporates techniques from Natural Language Processing (NLP), Information Retrieval (IR), and Artificial Intelligence (AI), has significant potential to reduce discovery time in medical and health research fields. In this talk, I will discuss how a self-learning based framework for knowledge discovery based on word embeddings can be designed to mine hidden associations between non-interacting medical concepts by rationally connecting independent nuggets of published literature. The self-learning process can model the evolutionary behavior of concepts to uncover latent associations between medical concepts, which allows us to learn the evolutionary trajectories of medical terms and detect informative terms in a completely unsupervised manner. Hence, meaningful hypotheses can be efficiently generated without prior knowledge. With the capability to discern reliable information from various resources, this self-learning framework provides a platform for combining heterogeneous resources and intelligently learning new knowledge with no user intervention.

Past UCI Seminars

May 8, 2017

Ram D. Sriram
“Information Technology for the Health Care Enterprise” by
Ram D. Sriram, National Institute of Standards and Technology
11 a.m. in Donald Bren Hall 3011

According to a recent report by the Center for Medicare and Medicaid Services (CMS), the United States spent nearly $3.2 trillion dollars on health care in 2015, which is about 18% of the nation’s GDP. Several studies have pointed out that properly implemented health information technology (HIT) could result in significant savings and improved health care. We use the term “health care informatics” for all software aspects of the healthcare enterprise: health information technology, modeling and simulation, bioinformatics, medical devices integration, and bioimaging. In this talk, I will discuss our work (at NIST) on health care informatics, focusing on: 1) Testing the nationwide health information network; 2) Medical device interoperability; 3) Systems biology/medicine, with a focus on protein-protein interaction; and 4) Medical/bioimaging.

April 21, 2017

Professor Wen Gao profile pictureComputer Science Seminar Series
11 a.m.- 12 p.m. in Donald Bren Hall 6011
“Multimedia Big Data based Personalized Health Management” by
Professor Wen Gao, Peking University
Host: Bren Professor Ramesh Jain

With the fact that “Healthy China” rises to national strategy, the medical services have transferred from after-disease treatment to preventive health management. Health management models vary from person to person, and the personalized health management needs to focus on the monitoring and analysis of individual lifestyle and behavior patterns. The real-time behavior data, which is automatically collected via wearable devices and social media, enables all-round recording of individual lifestyle and behavior patterns and thus can be exploited for personalized health management. There are some key issues we need to solve before building up this kind of system. First, how to storage and manage the huge personal multimedia data from surrounding camera and microphone and other devices. Second, how to acquire the personal historic data from the wearable devices they might keep all user data in their own loop. Third, quite close to the second issue but even hard to deal with, how we can get the real-time behavior data for the decision system of healthcare from wearable devices providers in case their data is not open to public yet. In this talk, I will discuss some solutions to solve the issues, including but not limited to, heterogeneous and multi-source health data structuring, standardization for the data format of wearable device, public multimedia data center for personal healthcare, disease association pattern mining from dynamic and static health data, personalized and real-time health management for dynamic decision support.