Speaker: Augusto Marques Ferreira Silva is an Associate Professor in the Department of Electronics, Telecommunications, and Informatics and a member of the Institute of Electronics and Informatics Engineering of Aveiro – IEETA. He is also the Director of the Master’s program in Medical Image Technologies. His preferred research area is Medical Image Processing with a focus on Machine Learning methods.
Abstract: In an era where data-driven decision-making has become central to healthcare, machine learning (ML) holds great promise. This presentation aims to promote discussion on the often overlooked but critical aspect of ML in healthcare: the pitfalls and challenges.
Key themes will include data quality and bias, interpretability, generalization, and ethical considerations. We’ll discuss how the unique characteristics of healthcare data, such as class imbalance and noisy labels, can skew model performance and decision-making.
Furthermore, we will emphasize the importance of interpretability and transparency in healthcare ML models. The ability to explain and understand model predictions is paramount when lives are at stake.
Generalization, or the ability of models to perform well on unseen data, is another critical challenge in healthcare. We will discuss strategies to mitigate overfitting and ensure that ML models perform reliably across diverse patient populations.
The role of latest trends such as generative ML will be addressed with some practical examples.
Ethical considerations, such as patient privacy, consent, and fairness, will be central to our discussion. We will address the ethical dilemmas that arise when deploying ML in healthcare and offer guidance on responsible AI implementation.
Location and date: IEETA auditorium, 15th November 2023, 15:30 Zoom link: https://videoconf-colibri.zoom.us/j/96216735526
