AAG 2025 Talks

AAG 2025 brought together several threads of my recent research: multimodal large language models, street view imagery, and multisensory urban experience. One presentation focused on using ChatGPT and Mapillary images to extract building attributes, while another explored the relationship between smell and vision in urban environments.

I like this pairing because the two projects ask related questions from different directions. The building attributes work looks at whether MLLMs can turn visual evidence into structured urban data. The multisensory work asks how visual environments connect to smell expectations and lived sensory experience. Together, they point toward a more flexible urban analytics toolkit: one that can read images, text, and human perception together.

Links: Connecting Senses project · Building Attributes project · Talks page