# Bayesian Modeling of Zero-Shot Classifications for Urban Flood Detection

arXiv: https://arxiv.org/abs/2503.14754 · HTML: https://mlsanity.com/paper/2503.14754/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 116 references that are in the map.
- First posted on arXiv 18 Mar 2025.

## Builds on (most cited first, 19 of 19)

- [Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning](https://mlsanity.com/paper/1801.07606/) (2018)
- [Language Models with Conformal Factuality Guarantees](https://mlsanity.com/paper/2402.10978/) (2024)
- [GPT-4V(ision) as a Generalist Evaluator for Vision-Language Tasks](https://mlsanity.com/paper/2311.01361/) (2023)
- [Large language model validity via enhanced conformal prediction methods](https://mlsanity.com/paper/2406.09714/) (2024)
- [Using Deep Learning and Google Street View to Estimate the Demographic Makeup of the US](https://mlsanity.com/paper/1702.06683/) (2017)
- [Learning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization](https://mlsanity.com/paper/2302.00275/) (2023)
- [Flood Prediction Using Machine Learning Models: Literature Review](https://mlsanity.com/paper/1908.02781/) (2019)
- [From Google Maps to a Fine-Grained Catalog of Street trees](https://mlsanity.com/paper/1910.02675/) (2019)
- [Detecting disparities in police deployments using dashcam data](https://mlsanity.com/paper/2305.15210/) (2023)
- [Fast Threshold Tests for Detecting Discrimination](https://mlsanity.com/paper/1702.08536/) (2017)
- [Domain constraints improve risk prediction when outcome data is missing](https://mlsanity.com/paper/2312.03878/) (2023)
- [Quantifying Spatial Under-reporting Disparities in Resident Crowdsourcing](https://mlsanity.com/paper/2204.08620/) (2022)
- [Equity in 311 Reporting: Understanding Socio-Spatial Differentials in the Propensity to Complain](https://mlsanity.com/paper/1710.02452/) (2017)
- [A Bayesian Spatial Model to Correct Under-Reporting in Urban Crowdsourcing](https://mlsanity.com/paper/2312.11754/) (2023)
- [Surveilling Surveillance: Estimating the Prevalence of Surveillance Cameras with Street View Data](https://mlsanity.com/paper/2105.01764/) (2021)
- [Redesigning Service Level Agreements: Equity and Efficiency in City Government Operations](https://mlsanity.com/paper/2410.14825/) (2024)
- [Evaluating multiple models using labeled and unlabeled data](https://mlsanity.com/paper/2501.11866/) (2025)
- [Social Sensing of Floods in the UK](https://mlsanity.com/paper/1711.04695/) (2017)
- [Fingerprinting New York City's Scaffolding Problem with Longitudinal Dashcam Data](https://mlsanity.com/paper/2402.06801/) (2024)

## Built on by (most cited first, 2 of 2)

- [Privacy of Groups in Dense Street Imagery](https://mlsanity.com/paper/2505.07085/) (2025)
- [Urban Incident Prediction with Graph Neural Networks: Integrating Government Ratings and Crowdsourced Reports](https://mlsanity.com/paper/2506.08740/) (2025)

## Abstract (arXiv)

Street scene datasets, collected from Street View or dashboard cameras, offer a promising means of detecting urban objects and incidents like street flooding. However, a major challenge in using these datasets is their lack of reliable labels: there are myriad types of incidents, many types occur rarely, and ground-truth measures of where incidents occur are lacking. Here, we propose BayFlood, a two-stage approach which circumvents this difficulty. First, we perform zero-shot classification of where incidents occur using a pretrained vision-language model (VLM). Second, we fit a spatial Bayesian model on the VLM classifications. The zero-shot approach avoids the need to annotate large training sets, and the Bayesian model provides frequent desiderata in urban settings - principled measures of uncertainty, smoothing across locations, and incorporation of external data like stormwater accumulation zones. We comprehensively validate this two-stage approach, showing that VLMs provide strong zero-shot signal for floods across multiple cities and time periods, the Bayesian model improves out-of-sample prediction relative to baseline methods, and our inferred flood risk correlates with known external predictors of risk. Having validated our approach, we show it can be used to improve urban flood detection: our analysis reveals 113,738 people who are at high risk of flooding overlooked by current methods, identifies demographic biases in existing methods, and suggests locations for new flood sensors. More broadly, our results showcase how Bayesian modeling of zero-shot LM annotations represents a promising paradigm because it avoids the need to collect large labeled datasets and leverages the power of foundation models while providing the expressiveness and uncertainty quantification of Bayesian models.