AI Tool Predicts Road Crashes: Johns Hopkins' SafeTraffic Copilot
Researchers at Johns Hopkins University have created an AI-based tool, Microsoft Copilot, to identify road crash risk factors and predict future incidents. Led by Hao 'Frank' Yang, an assistant professor, the tool aims to reduce traffic fatalities and injuries in the U.S.
Microsoft Copilot, published in Nature Communications, uses large language models to process and learn from vast amounts of data, including road condition descriptions, numerical values, and visual data. It improves prediction accuracy over time through a continuous learning loop and can quantify the trustworthiness of its predictions.
The model can evaluate individual and combined factor meals, providing a detailed understanding of how they influence crashes. It is designed to be a copilot for human decision-making, processing information and quantifying risks while humans remain the final decision-makers. The team, including Yang Zhao, Pu Wang, Yibo Zhao, and Hongru Du, plans to continue research on responsibly integrating AI-based models into high-stakes fields like public health and human safety.
Microsoft Copilot gives policymakers and transportation designers a trustworthy and interpretable tool to identify crash risk factors and execute evidence-based interventions. With improved prediction accuracy over time, it has the potential to significantly enhance road safety in the U.S.
Read also:
- Reconsidering the Approach to Mountain Height Measurement?
- UK automaker, Jaguar Land Rover, to commit £500 million for electric vehicle manufacturing in Merseyside
- Standard Nuclear & Framatome Join Forces to Boost TRISO Fuel Production by 2027
- Exhibition Spotlights Child Labor in Lithium and Cobalt Mines
