Raymond Tong - The Chinese University of Hong Kong
Effective motor recovery after stroke depends on intensive voluntary practice of the paretic limbs. We investigate the characteristics of brain waves and muscle activities related to the paretic upper limb movements after stroke. Brain waves and muscle activities are recorded by electroencephalography (EEG) and electromyography (EMG) respectively through electrodes attached to the scalp and skin surface. EEG has been applied in Brain-computer-interface (BCI), which is opening doors to use the power of the mind to overcome the limitations of the body. Then, we developed interactive control strategies to control different rehabilitation training systems for hand training in clinical trials, such as functional electrical stimulation (FES) and rehabilitation robot. The intelligence system incorporated the EMG and EEG as the bio-parameters to indicate the voluntary effort from a subject. We applied these engineering-based technologies in the field of Neurorehabilitation, robotic system uses electric motor to provide external assistive force during the rehabilitation training. The clinical studies showed functional improvement in the clinical outcome measures on the upper limb. The EEG and MRI analysis had shown the neuroplastic changes.
Wen J. Li - City University of Hong Kong
AI-based Sphygmographic Pulse Pattern Analysis for Traditional Chinese Medicine Palpation Diagnostics
Without X-ray computed tomography or stethoscope, Traditional Chinese Medicine (TCM) relies on inspection, auscultation/olfaction, inquiring, and palpation for medical diagnosis. We will present an intelligent “palpation robotic hand” (PRH) that uses 3 fingers, which are integrated with silicon and graphene-based pressure sensor arrays, to acquire temporal and 3D spatial pulse information from human wrists, and uses AI-based algorithms to classify arterial pulse patterns of patients. The technical specification of the PRH is to digitize and recognize at least 28 fundamental types of sphygmographic pulse patterns described by TCM doctors in the past few hundred years, and correlate these patterns to disease diagnosis. We will ultimately create a standardized database for TCM palpation pulse waves -- similar to the MIT-BIH Arrhythmia Database for electrocardiograph.
Wen J. Li is Chair Professor of Biomedical Engineering at the City University of Hong Kong (CityU). He was with The Chinese University of Hong Kong (1997-2011), NASA/CalTech Jet Propulsion Laboratory (1995-1997), and The Aerospace Corporation (1987-1994) before joining CityU. His academic honors include IEEE Fellow, ASME Fellow, and 100 Talents of the Chinese Academy of Sciences (中科院百人計劃). He served as the President of the IEEE Nanotechnology Council (2016 and 2017) and is currently VP of Academic Affairs (honorary) of the Shenzhen Academy of Robotics. He has co-founded Sengital Ltd. (Hong Kong) and Bewis Sensing LLC. (China), which are commercializing MEMS-based sensing systems worldwide. Prof. Li was educated at the University of Southern California (BSAE ‘87; MSAE ‘89) and the University of California, Los Angeles (PhD ‘97, Aerospace Engineering). His current research interest includes intelligent cyber physical sensors and AI for biomedical applications.
Raymond Louie - Hong Kong University of Science and Technology
Inferring the Fitness Landscape of HIV by Unsupervised Learning
An effective vaccine for HIV is not available, although hope has emerged through the discovery of antibodies capable of neutralizing diverse HIV strains. An ideal vaccine would elicit broadly neutralizing antibodies that target parts of the virus’s spike proteins where mutations severely compromise the virus’s fitness. Here, we employ an unsupervised learning approach that allows estimation of the fitness landscape (fitness as a function of sequence) of the polyprotein that comprises HIV’s spike. We validate the inferred landscape through comparisons with diverse experimental measurements. The availability of this fitness landscape will aid the rational design of immunogens for effective vaccines.
Raymond Louie is an IAS Junior Fellow and Research Assistant Professor at the Hong Kong University of Science and Technology. He received a PhD in Electrical Engineering at the University of Sydney in 2010. His research interests include applying statistical and machine learning tools to analyze big data and wireless communication systems. His recent work involved applying machine learning tools to viral sequence data, to aid in the design of effective vaccines.
Yefeng Zheng - Tencent YouTu Lab
Tencent AIMIS: AI-Powered Diagnostic Medical Imaging System
Yefeng Zheng received B.E. and M.E. degrees from Tsinghua University, China, and a Ph.D. degree from University of Maryland, USA. He is Director of Medical AI at Tencent, leading a team working on medical image analysis. Before that, he was Principal Key Expert at Siemens Healthineers. He has published 100+ papers，invented 70+ patents，and received several awards (including Thomas A. Edison Patent Award and Techno-College Innovation Award of European Association for Cardio-Thoracic Surgery). He is a Fellow of American Institute of Medical and Biological Engineering and an Associate Editor of IEEE Journal of Biomedical and Health Informatics.
Wanpracha Art Chaovalitwongse - University of Arkansas
The Future of Machine Learning in Quantitative Medical Image Analysis:
Radiomics has emerged as a quantitive approach for feature generation/engineering for quantitative image analysis. Radiomics features, coupled with machine learning, have been applied to medical imaging research with some degree of success. Recently, deep learning has gained more tractions in machine learning and is poised to become a new approach not only for prediction modeling but also for feature generation/engineering. In this talk I will share experience in using radiomics features versus deep learning-based features to predict clinical outcome from medical images in several real life clinical settings.
Dr. W. Art Chaovalitwongse is Professor of Industrial Engineering, 21st Century Leadership Chair in Engineering, and Co-Director of the Institute of Advanced Data Analytics at the University of Arkansas. He previously held faculty positions at the University of Washington, Princeton University, and Rutgers University. He is the author or co-author of over 150 published papers in machine learning, data analytics and optimization with applications in healthcare and other fields. In addition to undertaking academic research, he has worked/consulted for a number of companies such as ExxonMobil, Cisco, as well as the Bank of Thailand and several government entities in Thailand.
Qi Dou - The Chinese University of Hong Kong
Qi Dou is a final year PhD student in Computer Science and Engineering, The Chinese University of Hong Kong. Before that, she received her Bachelor degree in Biomedical Engineering from Beihang University in 2014. She has won the Best Paper Awards of Medical Image Analysis-MICCAI in 2017, Medical Imaging and Augmented Reality in 2016 and CUHK International Doctoral Forum in 2016. She was nominated MICCAI Young Scientist Award in 2016. Her research interests include medical image analysis and machine learning. She has published 20+ papers in this area on topics of medical image segmentation, biomarker detection, computer-aided diagnosis, multi-tasking on small medical dataset and domain adaptation. She pioneers 3D CNN for volumetric medical image analysis and it has become state-of-the-art deep learning method to compute radiological data. She serves as the reviewer for top-tier journals and conferences, including IEEE-TMI, IEEE-CYB, IEEE-TBME, Medical Image Analysis, MICCAI, IJCAI, etc.
Jeremy Lee - National Tsing Hua University
Learning a Phenotype Representation for AI-assisted Leukemia Diagnosis using Deep Generative Model
The advancement in deriving powerful data-driven representation has been key to the recent success across machine intelligence tasks. This representation embeds information in a high dimensional feature vector capturing the essence of complex and non-linear structure of our data space. In this talk, we will demonstrate the effectiveness of representational learning in the context of AI-assisted diagnosis application. A phenotype representation is learned using a deep generative model. Then, jointly with the National Taiwan University Hospital, we developed an AI-assisted interpretation algorithm of leukemia on over 10,000 unique patient’s flow cytometry data – achieving a remarkable accuracy >0.9 AUC.
Chi-Chun Lee (Jeremy) is an Assistant Professor at the Electrical Engineering Department of the National Tsing Hua University (NTHU), Taiwan. He received his Ph.D. degree in Electrical Engineering from the University of Southern California, USA in 2012. He was a data scientist at id:a lab at ID Analytics in 2013. His research interests are in the algorithmic development for human-centered behavioral signal processing (BSP) and affective computing. He has been involved in multiple granted interdisciplinary research projects, including aspects on education, psychology, neuroscience, and clinical health applications, with a focus on deriving decision analytics using signal processing and machine learning.
Peng Su - The Chinese University of Hong Kong
Deep Learning for Cardiovascular Health Monitoring
Peng Su is an Electronic Engineering PhD student at The Chinese University of Hong Kong. His research concentrates on deep learning and machine learning, especially spatiotemporal modelling and prediction. Currently, he is developing deep learning methods that leverage data-driven techniques to solve challenging problems in healthcare. And he is a co-founder and director of AI startup Intelligence Sensing Limited (Hong Kong), which utilise deep learning technology to improve human health.
Kelvin Tsoi - The Chinese University of Hong Kong
Artificial Intelligence for Dementia Screening
Dementia is a global public health problem among the ageing population worldwide. There are many screening tests available for the early detection of dementia; most of them are paper-based and involve geometric drawing. Drawing a simple picture, such as a pair of interlocking pentagons, is useful to evaluate visuospatial functions for people with dementia. With advancement of technology, drawing behaviour can be captured in real-time with digital platforms. Data can be captured in every second, so the response time is a kind of measurement for brain function. In this talk, Professor Kelvin Tsoi will demonstrate his dementia screening platform and machine learning models to analyze the drawing behaviour of dementia subjects. Real patient drawing samples will be used to show their work.
Kelvin is a Digital Epidemiologist. He received his Bachelor degree in Statistics and Doctoral degree in Medical Sciences in the Chinese University of Hong Kong. His research interests cover global cancer trends and other health conditions under the ageing population. He is leading an interdisciplinary research team on Big Data research for healthcare. Blood pressure monitoring on cloud system, automated screening platform for dementia are the typical examples. He has published over 50 full scientific articles in the foremost journals. A recent paper to compare dementia screening was published in JAMA Internal Medicine.
Lawrence Wee - Allianz Asia-Pacific
Artificial Intelligence in Mental Health
Epidemiological reports have highlighted mental illness as the leading causes of disability adjusted life years (DALYs) worldwide, with estimated global costs at nearly $2.5T with a projected increase to over $6T by 2030. Mental illness alone will account for more than half of the projected total economic burden from non-communicable diseases over the next two decades. Despite the burgeoning human, social and economic burden of mental illness, healthcare systems worldwide remain inadequately prepared, particularly in the heterogenous societies of Asia. This presentation focuses on recent technological advances and pilot studies around the application of machine learning and other domains of artificial intelligence in mental health – suggesting innovative digital health models for disease prevention, early intervention and management.
Dr. Lawrence Wee is a data science practitioner and innovator in the health sciences. He is currently Chief Data Scientist at Allianz Asia-Pacific where he directs the development of data-driven innovation platforms for the APAC markets. He works closely with the different stakeholders of the industry, co-creating solutions in the intersection of fintech, insurtech and healthtech.
Prior to this role, he founded the healthcare and biomedical analytics laboratory at the Institute of Infocomm Research (I2R) in Singapore, and served as principal investigator for research initiatives encompassing genomics, precision medicine, digital health and health services. After the stint at I2R, he moved on to Zuellig Pharma where he took on the role of Chief Data Scientist.
Dr. Wee sits on the scientific advisory boards of several technology companies, and speaks regularly at conferences on artificial intelligence, digital technology and healthcare. He was previously the President of the Association for Medical and Bio-informatics Singapore (AMBIS), and the current vice-chair of the Chinese-American Pharmaceutical Society (CABS) Singapore chapter. He received his Ph.D. in Computational Biology and Machine Learning from the National University of Singapore.
PANEL DISCUSSION: From Innovation to Implementation in Healthcare
CONVERSATION AND DRINKS
Ta-Chu Kao - Unievrsity of Cambridge
Ta-Chu obtained his master degree in Physics and Philosophy at Oxford University. Currently, he is a PhD student in the Computational and Biological Learning Lab at Cambridge University. The focus of his PhD research is: How do neutrons work together to generate and control body movements? Applying control theory and machine learning methods, he builds biologically-realistic models of the motor cortex and develops new techniques for analysing large-scale neural recordings.
AI-Powered Doctor App and Remote Patient Monitoring
Wearable Healthcare Monitoring Sensors
Lowering Drug Discovery Costs with Machine Learning
Leveraging Machine Learning to Reduce Medical Errors
Artur Kadurin - Insilico Medicine
Longenesis: returning the control over personal and medical data back to the individual using AI-driven blockchain solutions
Longenesis is a revolutionary blockchain-based Life Data Marketplace platform which uses Artificial Intelligence (AI) to store, manage, and trade life data: medical records, social network, and health data. This platform facilitates human data transactions between Contributors(the general public) and Customers (Drug development/Pharmaceutical companies). Longenesis is a partnership by Bitfury and Insilico Medicine, two leaders within the Blockchain and Medical AI industries. Longenesis offers a solution to both problems: by operating on blockchain network we ensure advanced security and anonymity on the part of the users. Contributors will be compensated for every piece of information they choose to provide(ranging from a selfie to a full blood work file) by LifePound - cryptocurrency and the central monetary tool of the Longenesis Marketplace. This information is then captured, analyzed and stored by the integrated Deep Learning Medical AI developed by Insilco Medicine.
Artur Kadurin is Chief AI Officer at Insilico Medicine, and CEO Insilico Taiwan - companies utilizing Deep Learning techniques for drug discovery and aging research. He has strong mathematical background and experience in all flavors of Data Science, Machine Learning and Deep Learning. Prior to Insilico Medicine, he worked at Mail.ru Group - one of the biggest European IT company as the lead data scientist, where he built from scratch user segmentation system for biggest in Russia advertising system. His projects included predictive advertising analysis, social network analysis, ranking systems, etc. In collaboration with Insilico Medicine, he published a paper on the first application of GANs (Generative Adversarial Networks) techniques within the field of cancer drug discovery. At the moment he is leading cutting-edge research within the Pharma.AI division of Insilico Medicine.
Fei Jiang - The University of Hong Kong
AI powered clinical trial design
The goal of a typical phase II clinical trial design is to test the treatment effacy based on the responses on treatments from the study samples. We cast the problem in a reinfoncement learning context and develop a class of decision theoretic phase II clinical trial designs with sequential stopping and adaptive treatment allocation to evaluate treatment efficacy. The sequential stopping allows the trial to terminate early so as to minimize the patient enrollment in the trial. The adaptive radnomization allows the patients have high probability to be allocated to the promising treatment. We develop an algorithm based on the constrained backward induction and forward simulation to implement the designs. The algorithm overcomes the computational difficulty of the backward induction method, thereby making our approach practicable. The designs result in trials with desirable operating characteristics under the simulated settings. Moreover, the designs are robust with respect to the response rate of the control group.
Fei Jiang is an Assistant Professor in the Department of Statsitics and Actural Science, The University of Hong Kong. She obtained her Ph.D. from Rice University, U.S.A, 2013. Before join the University of Hong Kong, she was a postdoctoral fellow in Harvard University from 2013 to 2016. She is specialized in statsical machine learinng in health care research, including clinical trial design, epidemiologies, enviormental health, functional data analysis, image data analysis. Her work has been pubished on high rank statisticcal journals. Besides intensive reseasrch, she is actively collaborating with heath authorities and hospitals to implement the machine learning algirthms in the real medical context.
Ankur Purwar - Procter & Gamble
Diagnostics and Personalization in Skin Care via Artificial Intelligence
The mass beauty aisle is often crowded and confusing and experiences with beauty counselors in specialty department stores can be overwhelming. In both cases, women are walking away less than satisfied with their shopping experience and not certain that their products are exactly right for their skin. Part of this dissatisfaction stems from the fact that women want personalized attention and recommendations, and the expectation of what “personalized” means has changed. Dr Ankur Purwar will discuss more on the newly developed Olay Skin Advisor, a web-based skin analyst and advisor tool that uses AI to address this problem.
Dr. Ankur Purwar is a Principal Scientist in the Procter & Gamble Beauty Technology organization based in Singapore. A serial innovator with passion for transforming and designing experiences on how consumers interact with our products. His research is currently focused on linking perception and noticeability for appearance understanding leveraging computer vision, and machine learning techniques to measure skin and hair attributes. Ankur holds a PhD in Applied Mathematics with focus on Medical Imaging Analysis from IIT Kanpur, India.
Simon Poon - University of Sydney
Dr Simon K. Poon is an Associate Professor at the University of Sydney specialising in Health Informatics Research. He is a cross-disciplinary researcher with training in diverse disciplines including Computer and Mathematical Sciences, Management, Information Systems and Public Health. His research interests encompass the interaction of health and technology. With the increasing shift towards collaborative healthcare and the increase in the use of digital technologies, one of Simon’s key research agenda is to foster methodological advances in health technology evaluation.
Ethical Dilemmas & Regulation
END OF SUMMIT