Welcome to RWGAI (Real World Generalisation in AI), where we focus on advancing the robustness and domain generalisation of artificial intelligence systems. Our research is centred on developing AI models that exhibit strong performance across diverse, real-world environments, ensuring they can adapt effectively to previously unseen tasks and data distributions. In addition to our cutting-edge research, we provide comprehensive benchmarks and datasets to support the evaluation and development of generalisable AI systems. By prioritising robustness and enabling true domain generalisation, RWGAI strives to push the boundaries of AI, facilitating the deployment of intelligent systems that perform reliably across a wide range of practical applications.
Datasets and Benchmarks
Recent Publications
- Accepted: Sara Al-Emadi, Yin Yang, Ferda Ofli, Benchmarking Object Detectors under Real-World Distribution Shifts in Satellite Imagery, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2025, Nashville, United States.