Urban Safety Perception Through the Lens of Large Multimodal Models
Abstract: Understanding how urban environments are perceived in terms of safety is crucial for urban planning and policymaking. Traditional methods like surveys are limited by high cost, required time, and scalability issues. To overcome these challenges, this study introduces Large Multimodal Models (LMMs), specifically Llava 1.6 7B, as a novel approach to assess safety perceptions of urban spaces using street-view images. In addition, the research investigated how this task is affected by different socio-demographic perspectives, simulated by the model through Persona-based prompts. The model achieved an average F1-score of 59.21% in classifying urban scenarios as safe or unsafe. Incorporating Persona-based prompts revealed significant variations in safety perceptions across the socio-demographic groups of age, gender, and nationality. Elder and female Personas consistently perceive higher levels of unsafety than younger or male Personas. Similarly, nationality-specific differences were evident in the proportion of unsafe classifications ranging from 19.71% in Singapore to 40.15% in Botswana. Notably, the model’s default configuration aligned most closely with a middle-aged, male Persona. Interestingly, model's decisions are based on visual features like isolation and lack of maintenance that are in line with multiple urban social theories like Eyes on the Street and Defensible Space. Our findings highlight the potential of LMMs as a scalable and cost-effective alternative to traditional methods for urban safety perceptions. While the sensitivity of these models to socio-demographic factors underscores the need for thoughtful deployment, their ability to provide nuanced perspectives makes them a promising tool for AI-driven urban planning.
Short bio: Dr. Massimiliano Luca is a senior researcher at the Mobile and Social Computing Lab - Center for Augmented Intelligence - Fondazione Bruno Kessler, Trento, and a member of the Complex System Society Council. In November 2023, he obtained his PhD at the Free University of Bolzano/Bozen and, during his PhD he had the opportunity to be a Visitor Researcher at MIT, Boston and at the Spanish National Research Council (CSIC). He authored more than 20 research papers focused on artificial intelligence applied to mobility, transportation and city science. Such publications received more than 5 awards from different conferences and journals. Previously, Massimiliano was the Chief Scientific Officer at Pulse.io, a London-based start-up that employs artificial intelligence in emotional marketing where he was responsible for data science, web-services implementation, cloud computing, deep learning and generative AI assets.
Presenter: Dr. Massimiliano Luca (Bruno Kessler Foundation, Trento, Italy)
Date: 2025-04-09 12:15 (CEST)
Location: Oficinas ELLIS Alicante, Muelle Pte., 5 – Edificio A, Alicante 03001, Alicante ES
Add to Google Calendar, Outlook Calendar
This talk is part of the ELLIS / FBK-MobS Workshop.