November 17, 2022 16:06
TrustML Young Scientist Seminar #41 20221116 thumbnails

Description

The 41st Seminar
Date and Time: Nov. 16th 4:30 pm – 5:30 pm(JST)
Venue: Zoom webinar
Language: English

Speaker: Phi Le Nguyen (Hanoi University of Science and Technology)
Title: Trustworthy AI in SmartHealth and a case-study in Vietnam
Short Abstract:
Vietnam is in a severe shortage of physicians, with the ratio of doctors and nurses per population being much lower than the average of low- and middle-income countries. This situation necessitates the development of systems that help Vietnamese be proactive in taking care of their health, monitoring risks in the pre-disease stages, and thereby improving healthcare quality in general. One obvious solution is to digitize healthcare information and deliver it to every citizen. The VAIPE project aims to build an intelligent healthcare system to assist users in collecting, managing, and analyzing their health-related data. Our system enables users to collect heterogeneous data captured from multiple sources using a convenient smartphone camera, provides visualizations of analytical and predicted results, and includes functions to support users, for example, reminding of medication schedules and warning of early-disease risks. Our system is AI-assisted and involves original research and development of several key modules: (1) representation, storage, and processing of multi-source multi-type data, (2) training, learning, and mining on data for clinical insights and disease risk prediction with supporting evidence, (3) enhancement of user privacy and engagement in sharing their health-related data, and (4) optimized resource allocation to reduce deployment cost while guaranteeing QoS constraints. In this talk, I would like to share our recent findings on trustworthy AI in VAIPE. Specifically, I will focus on pill detection with reliability and explainability, and Federated learning under non-ideal and uncontrollable clients’ data.