Humanitarian AI

AIDR – Artificial Intelligence for Digital Response

AIDR—the Grand Prize winner of the 2015 Open Source Software System Challenge—is a free and open platform to filter and classify social media messages related to emergencies, disasters, and humanitarian crises. AIDR uses human and machine intelligence to automatically tag up to thousands of messages per minute. https://aidr.qcri.org

DisasterVQA

Social media imagery offers a low-latency source of situational information during natural and human-induced disasters, but the complex, safety-critical reasoning required for disaster response is poorly served by general-purpose vision–language models. We introduce DisasterVQA, a benchmark dataset designed for perception and reasoning in crisis contexts. It comprises 1,395 real-world images and 4,405 expert-curated question–answer pairs spanning diverse events such as floods, wildfires, and earthquakes. Grounded in established humanitarian frameworks including FEMA’s Emergency Support Functions (ESF) and OCHA’s Multi-Cluster/Sector Initial Rapid Assessment (MIRA), the dataset features binary, multiple-choice, and open-ended questions that target both situational awareness and operational decision-making. DisasterVQA provides a challenging and practical benchmark to guide the development of more robust and operationally meaningful vision–language models for disaster response.

DSGR

We introduce Domain Shift across Geographic Regions (DSGR), a new large-scale dataset designed to study the effects of real-world geospatial distribution shifts in satellite imagery classification. DSGR captures variability across diverse geographic regions, with particular emphasis on underrepresented areas such as Africa and Oceania, enabling systematic analysis of how regional differences impact model performance. This dataset is motivated by a fundamental limitation of deep learning models: they are typically trained under the i.i.d. assumption and often suffer significant performance degradation when deployed in environments that differ from the training data. Domain Generalisation (DG) aims to address this challenge by improving model robustness to Out-Of-Distribution data without access to target domains during training. By explicitly modelling geographic domain shifts, DSGR provides a valuable benchmark for advancing DG research in satellite imagery classification.

Flood Insights

The Flood Insights Dashboard (FID) is a comprehensive tool for evaluating the immediate flood situation during the onset of a flood disaster. Using the best available data, methods and analytical tools, FID synthesizes satellite, social media and geospatial data to help guide crisis responders in making effective decisions.

Global Landslide Detector

The development of a system that monitors social media continuously for general landslide-related content using a landslide classification model to identify and retain the most relevant information is described and validated. The system harvests photographs in real-time from these data and tags each image as landslide or not-landslide. A training model was developed with input from computer scientists, geologists (landslide specialists) and social media specialists to establish a large image dataset that has then been applied to the live Twitter data stream. The preliminary model was developed by training a convolutional neural network on the dataset. Quantitative verification of the system’s performance during a real-world deployment shows that the system can detect landslide reports with Precision = 76%.

Incidents1M

Natural disasters, such as floods, tornadoes, or wildfires, are increasingly pervasive as the Earth undergoes global warming. It is difficult to predict when and where an incident will occur, so timely emergency response is critical to saving the lives of those endangered by destructive events. Fortunately, technology can play a role in these situations. Social media posts can be used as a low-latency data source to understand the progression and aftermath of a disaster, yet parsing this data is tedious without automated methods. Prior work has mostly focused on text-based filtering, yet image and video-based filtering remains largely unexplored. In this work, we present the Incidents1M Dataset, a large-scale multi-label dataset which contains 977,088 images, with 43 incident and 49 place categories. We provide details of the dataset construction, statistics and potential biases; introduce and train a model for incident detection; and perform image-filtering experiments on millions of images on Flickr and Twitter. We also present some applications on incident analysis to encourage and enable future work in computer vision for humanitarian aid.

Mapping Education Insecurity

This visual shows reports of attacks on education in Africa and the Middle East from the Armed Conflict Location & Event Data Project (ACLED) as well as social media posts from Twitter about education insecurity identified by the Artificial Intelligence for Digital Response (AIDR) platform. The blue circles represent the number of tweets about attacks on education; the orange triangles represent verified reports of attacks on education from ACLED.

RWDS

Object detectors achieve strong performance on benchmark datasets, yet most are trained under the i.i.d. assumption, leading to significant degradation when deployed under real-world distribution shifts. Domain Generalisation (DG) addresses this challenge by enabling models to generalise to unseen, Out-Of-Distribution data without access to target domains during training. However, evaluating object detection under realistic DG conditions remains difficult due to the lack of standardised benchmarks. To fill this gap, we introduce Real-World Distribution Shifts (RWDS), a suite of three benchmark datasets designed to assess the robustness of state-of-the-art object detectors under realistic spatial domain shifts. Grounded in humanitarian and climate change applications, RWDS enables systematic evaluation across diverse climate zones, disaster types, and geographic regions, providing a timely benchmark for developing object detectors that generalise beyond the training setup.

Multimodal (text + image) Twitter dataset from seven major natural disasters, annotated for humanitarian response tasks.

Human-annotated datasets, lexicons, and word embeddings for crisis-related social media analysis.

DSGR

DSGR - Analysing Satellite Imagery Classification under Spatial Domain Shift across Geographic Regions Abstract Deep learning models are designed based on the i.i.d. assumption; consequently, they experience a significant performance drop due to the distribution shifts when deployed in real environments. Domain Generalisation (DG) aims to bridge the distribution shift between the source and target domains by improving the generalisability of the model to Out-Of-Distribution (OOD) data. This challenge is prominent in satellite imagery classification due to the scarcity of data from underrepresented regions such as Africa and Oceania. In this paper, we address the limitations of existing datasets in capturing distribution shifts caused by geospatial differences between geographic regions by constructing a new, large-scale dataset called Domain Shift across Geographic Regions (DSGR). This dataset aims to help researchers better understand the impact of distribution shifts on satellite imagery classification. Furthermore, we perform rigorous experiments on DSGR to investigate and benchmark the robustness of existing DG techniques under single- and multi-source domain settings and the role of foundation models in enhancing the DG techniques. Our evaluations reveal that recent DG techniques have a comparable, yet weak, performance on DSGR. However, when combined with a foundation model like CLIP, ERM (introduced in 1999) achieves highly competitive results, surpassing even recent state-of-the-art DG solutions in enhancing the generalisability of deep learning models across different geographic regions. Our dataset and code are available at https://github.com/RWGAI/DSGR.

RWDS

RWDS - Benchmarking Object Detectors under Real-World Distribution Shifts in Satellite Imagery (CVPR 2025) Abstract Object detectors have achieved remarkable performance in many applications; however, these deep learning models are typically designed under the iid assumption, meaning they are trained and evaluated on data sampled from the same (source) distribution. In real-world deployment, however, target distributions often differ from source data, leading to substantial performance degradation. Domain Generalisation (DG) seeks to bridge this gap by enabling models to generalise to Out-Of-Distribution (OOD) data without access to target distributions during training, enhancing robustness to unseen conditions. In this work, we examine the generalisability and robustness of state-of-the-art object detectors under real-world distribution shifts, focusing particularly on spatial domain shifts. Despite the need, a standardised benchmark dataset specifically designed for assessing object detection under realistic DG scenarios is currently lacking. To address this, we introduce Real-World Distribution Shifts (RWDS), a suite of three novel DG benchmarking datasets that focus on humanitarian and climate change applications. These datasets enable the investigation of domain shifts across (i) climate zones and (ii) various disasters and geographic regions. To our knowledge, these are the first DG benchmarking datasets tailored for object detection in real-world, high-impact contexts. We aim for these datasets to serve as valuable resources for evaluating the robustness and generalisation of future object detection models. Our datasets and code are available at https://github. com/RWGAI/RWDS.