Dr Xiaochen Yang she/her

My research passion lies in understanding the mechanisms behind deep learning using mathematical tools – in particular, unveiling the ‘black box’ of neural networks to improve their data and training efficiency, as well as their robustness. This line of work not only advances fundamental machine learning theory but also enables impactful real-world applications, especially in medical science. I am particularly interested in applying trustworthy and data-efficient AI to healthcare challenges, such as the early diagnosis of neurodegenerative diseases and improving the well-being of affected patients. In my research group, we develop methods in trustworthy AI – including robustness to adversarial attacks – and few-shot learning, with recent work focusing on the fine-tuning of vision-language models (VLMs). On the applied side, we have developed data-efficient and interpretable AI techniques for Alzheimer’s diagnosis based on brain imaging. I am keen to expand collaborations with healthcare researchers, clinicians, and potentially those developing smart sensors, to ensure our tools are robust, practical, and clinically useful.

Interdisciplinarity is a key aspect of my current and future work. I am already involved in projects aimed at improving the well-being of older adults, and I would like to develop CDT projects that integrate perspectives from computer science, medicine, psychology, public health, and engineering. As a supervisor, I offer close guidance early on and then support students to grow into independent researchers. I encourage collaboration, participation in research events, and wider academic engagement.

I am committed to equity, diversity and inclusion in research and teaching. I have co-organised several widening participation and outreach activities to encourage school pupils from underrepresented backgrounds to study mathematics and pursue university education.
Outside of work, I enjoy hiking and scuba diving, which offer perspective and spark creativity. I look forward to contributing to the DiveIn CDT community, whose mission-driven, inclusive, and interdisciplinary vision strongly aligns with my own research values.