UGOD Seminar | Towards more Explainable Urban Design Analytics
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
As AI and urban analytics become increasingly ubiquitous in urban design, it is important to move beyond prediction towards explanation. This talk explores the importance of explainable urban design analytics through two case studies: generative counterfactuals of street views, and street counterfactuals. Together, they demonstrate how explainability methods can make AI more interpretable. Despite their usefulness, the underlying models remain, to a large extent, opaque. Explainability should therefore be seen not as a solution, but as a call towards more critically informed use of AI in urban design analytics.
Dr Stephen Law’s research interests centre on leveraging urban and spatial data science to plan more sustainable and liveable cities. He holds a Bachelor’s in Economics, a Master’s in Urban Design, a PhD in UCL Bartlett Space Syntax Lab, a postdoctoral fellowship at the Alan Turing Institute, Associate Professorship in UCL(Geography) and have recently taken up an associate professorship role at CUHK in the coming school year. Dr Law has authored peer-reviewed publications and secured multiple external grant funding, including research on housing deprivation, urban health, disaster planning and spatial cognition.