New mathematical tools shed light on the fluctuations of living matter

Fluctuations in such energy-consuming systems cannot be assessed by traditional physics due to the influence of the arrow of time on their behavior

Quantitative predictions on the behavior of active matter can facilitate the experimental design of such systems

Chalmers University of Technology, Sweden, has received 13 million SEK from the Hasselblad Foundation to build a new camera for Sweden’s premier radio telescope. The new three-band camera is part of a global initiative to enable high-quality, multi-colour images of black holes and other space phenomena, using innovative technology that has excited astronomers since its first test runs in South Korea.

To answer this question: Can generative AI improve vehicle trajectory prediction in car-following scenarios? Researchers from the University of Wisconsin–Madison, Tongji University, and collaborators developed FollowGen, a conditional diffusion model that integrates historical motion features and inter-vehicle interactions to generate safer and more reliable trajectory predictions for autonomous driving.

Cross-city transfer learning (CCTL) has emerged as a crucial approach for managing the growing complexity of urban data and addressing the challenges posed by rapid urbanization. This paper provides a comprehensive review of recent advances in CCTL, with a focus on its applications in urban computing tasks, including prediction, detection, and deployment. We examine the role of CCTL in facilitating policy adaptation and influencing behavioral change. Specifically, we provide a systematic overview of widely used datasets, including traffic sensor data, GPS trajectory data, online social network data, and map data. Furthermore, we conduct an in-depth analysis of methods and evaluation metrics employed across different CCTL-based urban computing tasks. Finally, we emphasize the potential of cross-city policy transfer in promoting low-carbon and sustainable urban development. This review aims to serve as a reference for future urban development research and promote the practical implementation of CCTL.

To answer this question: Can Traffic Accident Reports Aid Visual Accident Anticipation? A research team led by Professor Zhenning Li from the University of Macau proposes a visual-textual dual-branch traffic accident prediction framework that leverages domain knowledge, aiming to achieve high-performance, high-efficiency, and explainable accident anticipation.

Researchers at the University of Wisconsin–Madison have developed a control framework to enable safe and robust docking of Modular Autonomous Vehicles (MAVs) under uncertainty. The proposed method combines adaptive control with safety barrier functions and is validated through both simulation and the first-ever field test of MAV docking using a reduced-scale robotic platform.

In this study, we proposed a novel Knowledge-Informed Deep Learning (KIDL) paradigm that, to the best of our knowledge, is the first to unify behavioral generalization and traffic flow stability by systematically integrating high-level knowledge distillation from LLMs with physically grounded stability constraints in car-following modeling. Generalization is enhanced by distilling car-following knowledge from LLMs into a lightweight and efficient neural network, while local and string stability are achieved by embedding physically grounded constraints into the distillation process. Experimental results on real-world traffic datasets validate the effectiveness of the KIDL paradigm, showing its ability to replicate and even surpass the LLM's generalization performance. It also outperforms traditional physics-based, data-driven, and hybrid CFMs by at least 10.18% in terms of trajectory simulation error RMSE. Furthermore, the resulting KIDL model is proven through theoretical and numerical analysis to ensure local and string stability at all equilibrium states, offering a strong foundation for advancing AV technologies.

Practically, KIDL offers a deployable solution for AV control, serving as a high-level motion reference that ensures realistic and stable car-following in mixed traffic environments. Moreover, this framework provides a promising pathway for integrating LLM-derived knowledge into traffic modeling by distilling it into a lightweight model with embedded physical constraints, balancing generalization with real-world feasibility.