3 accepted papers
Effective time series representation is critical for revealing temporal dynamics in many fields. However, existing approaches encounter fundamental limitations. Discrete-time representations struggle with irregular sampling and the tradeoff of fidelity and efficiency, while traditional implicit neur…
Unsupervised anomaly detection of multivariate time series remains challenging in complex nonstationary dynamics, due to the high false-positive rates and limited interpretability. We propose PhysDiff, combining physics-guided decomposition with diffusion-based reconstruction, to address these issue…
Several microrobots driven by capacitive actuators have shown excellent performance because of the high power density and low working current of the actuators. However, these capacitive actuators need high alternating working voltage (hundreds or thousands of volts), which results in heavy and bulky