Building Trustworthy AI Models for Clinical Environments
Driven by a passion for translating Artificial Intelligence (AI) research into practical solutions, Dr Fu Huazhu collaborates with fellow researchers and industry partners to transform medical diagnostics.
Artificial Intelligence (AI) is transforming medical diagnostics, enabling faster and more precise analysis of medical images to support clinical decision-making. However, AI models are only as reliable as the data on which they are trained.
In real-world clinical settings, the quality of medical images can vary significantly, introducing uncertainty. Without careful verification by trained human interpreters, AI models risk generating incorrect medical information, leading to serious misdiagnosis and treatment errors.
Dr Fu Huazhu is tackling this challenge head-on, ensuring AI-driven diagnostics remain accurate and trustworthy.
Dr Fu and his team seek to identify and quantify such uncertainty in medical image analysis. His work is not just theoretical; he applies it to real-life settings such as diagnosing eye diseases, where he works with ophthalmologists to develop more robust methods for detecting eye anomalies in a clinical environment.
The complexity of the real-world data means that we need to design uncertainty into these predictive AI models. By doing so, we are able to flag problematic data to its human operator, thus improving its ‘trustworthiness’ and reliability.
As a Principal Scientist at A*STAR Institute of High Performance Computing (A*STAR IHPC), Dr Fu has spent years pushing the boundaries of AI, and has published over 300 widely cited academic articles. Whether it's enhancing healthcare systems, optimising transport networks, or transforming manufacturing processes, Dr Fu thrives on collaborating with industry partners to turn AI models into practical solutions. For him, science is not just about discovery; it's about making a difference.
True to his passion for finding solutions to real-world problems, Dr Fu looks for the same drive and enthusiasm in the young researchers he recruits and nurtures.
“While they need to possess the technical skills in programming, I also look for those who truly enjoy research, are keen to learn and are open to collaboration with others,” he shares.
“Although we focus on healthcare and biomedical science, AI applies across nearly all industries and areas. You just need to be willing to dive deep enough into the process of finding a solution.”
