Peter W
- Research Program Mentor
PhD at Stanford University
Expertise
machine learning for healthcare, mobile and web development for healthcare
Bio
I am an Assistant Professor in Computer Science at the University of Hawaii at Manoa. I completed my PhD and MS at Stanford University and BA at Rice University. My research interests are in creating scalable, accessible, and usable digital health solutions for various psychiatric, developmental, and mental health conditions. My project interests include digital diagnostics of conditions and adaptive digital therapies for conditions using a combination of web/mobile applications and machine learning. Outside of research, I naturally enjoy going to the beach given that I am a professor in Hawaii. Besides that, I enjoy raising my dog with my wife and playing/making music. I am looking for students with either web development, mobile development, data analysis, or machine learning experience.Project ideas
Machine Learning for Human Behavior
There are many human behaviors which can be classified from data streams such as video, audio, smartphone usage data, and biosignals. Using out-of-the-box machine learning methods usually do not cut it for classifying complex human behavior, especially those relevant to psychiatry and mental health. This project involves developing a novel innovation in data augmentation, pre-training, or representation learning to improve an AI model for either emotion recognition, activity detection, stress/anxiety prediction, and other behaviors with publicly available datasets. The resulting research paper would compare the classification for the human behavior with and without your innovation.
Digital Diagnostics
Digital diagnostics revolutionize the field of healthcare by leveraging advanced technologies to transform the way medical conditions are diagnosed. Combining artificial intelligence, machine learning, and big data analytics, digital diagnostics offer faster, more accurate, and cost-effective methods of assessing and monitoring patients' health. These innovative tools enable healthcare professionals to gather comprehensive patient data, including symptoms, medical history, and vital signs, and process it efficiently to generate precise diagnostic insights. By harnessing the power of digital platforms, such as mobile apps, wearables, and telemedicine, digital diagnostics empower individuals to actively participate in their healthcare, promoting early detection, personalized treatment plans, and improved patient outcomes. With the potential to revolutionize healthcare delivery, digital diagnostics represent a significant milestone in the quest for more efficient and patient-centric medical practices. This project involves developing a smartphone app or website which can be used to diagnose 1 or more health conditions. Prior projects have included Autism, Parkinson's, Dementia, Hypertension, Scoliosis, Alzheimer's, and Depression.
Digital Therapeutics
Digital therapeutics are a groundbreaking approach to healthcare that harnesses the power of technology to deliver evidence-based interventions for the prevention, management, and treatment of various medical conditions. These innovative solutions leverage software programs, mobile apps, wearable devices, and virtual platforms to provide personalized, scalable, and accessible treatments. Digital therapeutics go beyond traditional pharmaceutical interventions, offering interactive and engaging experiences that empower individuals to take an active role in their own health. Through cognitive behavioral therapy, mindfulness exercises, medication adherence reminders, and personalized health coaching, digital therapeutics aim to improve outcomes, enhance patient engagement, and reduce healthcare costs. These digital interventions are backed by rigorous scientific research, making them a promising addition to the healthcare landscape, with the potential to augment traditional approaches and transform the way we prevent, manage, and treat diseases. This project involves developing a smartphone app or website which can be used to help people with 1 or more health conditions.
Machine Learning on Electronic Health Records
Machine learning techniques can be used to analyze the vast wealth of electronic health records (EHRs) contained within the National Institutes of Health (NIH) All of Us dataset, which contains data on hundreds of thousands of people. The goal of this project is to uncover valuable insights and patterns within the extensive collection of de-identified patient data, which spans diverse populations and health conditions. By leveraging the power of machine learning algorithms, researchers are able to sift through this treasure trove of information and identify hidden correlations, risk factors, and predictive models for various diseases and health outcomes. This project holds the potential to revolutionize healthcare by enabling precision medicine approaches, tailoring treatments, and interventions to individual patients based on their unique characteristics and medical histories. Furthermore, the findings derived from this research could inform policy decisions, enhance healthcare delivery, and contribute to the advancement of population health management strategies. With its scale and scientific rigor, this machine learning-driven analysis of EHRs represents a significant step forward in unlocking the transformative potential of big data in healthcare research and improving the health and well-being of individuals on a global scale.
