To address the limitations of traditional Coriolis-force experiments—such as coarse measurements and single-function setups—this study develops an intelligent demonstration and investigation platform for Coriolis dynamics. The platform adopts an end-cloud collaborative distributed architecture: the front-end consists of an STM32 control unit and a K230 vision-acquisition module for multi-source data collection, while the back-end host computer integrates Bluetooth and Wi-Fi dual-channel data transmission. By incorporating a YOLOv8-based object detection model, the system enables real-time tracking of dynamic processes and simultaneous comparison between theoretical predictions and experimental measurements. Performance evaluations show that the platform achieves a spatial resolution of 0.667 mm/pixel, an end-to-end latency of less than 1.320s, and a trajectory-fit goodness（R²）above 0.99 between YOLO-extracted experimental trajectories and theoretical trajectories over 50 repeated trials under identical conditions, demonstrating high precision, low latency, and strong reliability.Through two representative experiments—ball motion deflection and fluid-erosion patterns—the platform is shown to support both intuitive visualization of classical physical laws and in-depth investigation of complex nonlinear phenomena. By integrating the Mamba model with Physics-Informed Neural Networks（PINN）, and further introducing semi-supervised learning and self-attention mechanisms, the platform is capable of predicting the outcomes of fluid erosion experiments and inferring the dominant physical mechanisms governing the erosion process directly from data. This work provides a complete and practical solution for the modernization and intelligent upgrading of traditional physics experimentation systems.

Data science has always been the driving force behind the dawn of the Fourth Industrial Revolution. Over the past decade, the sharp decline in the cost of sensors, data storage, and computing resources has made data-driven discovery methods possible, which has had a transformative impact on science and promoted various innovations in characterizing high-dimensional data generated by experimental observations. The field of AI4 Science aims to apply AI to physical domains. However, mainstream deep learning methods are criticized as a “black box,” with their internal workings being difficult to understand. Therefore, interpretable machine learning aims to break through the “black box” mechanism, allowing us to understand the internal workings of machine learning in a human-readable manner. This makes AI-assisted scientific discovery possible. This paper takes the damped oscillator as the research object. A YOLOv8 visual model trained on a custom dataset is used to obtain spatiotemporal data. On one hand, symbolic regression with the introduction of information criteria is employed to automatically discover explicit expressions between data. On the other hand, we conceptualize the problem of dynamical discovery from the perspective of sparse regression, constructing a candidate function library through numerical differentiation with total variation regularization, and ultimately accurately reconstructing the differential equation and nonlinear laws of damped vibration. We have verified the strong scientific interpretability and universality of the results obtained by this method using both experimental and simulation data.

This device was developed using a project-based learning approach, with the goal of accurately measuring weak geomagnetic fields. We independently designed a measuring instrument, starting with the Gaussian method, to conduct preliminary measurements of the horizontal component of the geomagnetic field. Then, we improved the Gaussian method using a self-made Helmholtz double coil. Finally, based on the communication between a LabVIEW host computer and an Arduino microcontroller, a photoelectric probe detection system was established, enabling efficient and convenient data transmission and processing, as well as higher measurement accuracy. Additionally, we independently designed a Hall-Gauss meter to measure the vertical component of the geomagnetic field using the flipping method. To adapt to the diversity of magnetic field measurement scenarios, we used a STM32 microcontroller to make a self-made AMR Gauss meter and independently wrote hard magnetic correction and tilt compensation algorithms to replace the Kalman filter algorithm. This instrument is characterized by low cost, good interactivity, portability, high accuracy, and can be applied to a wider range of application scenarios.