Time Scale
Time scale analysis focuses on understanding and modeling systems exhibiting dynamics across multiple temporal resolutions, aiming to improve prediction accuracy and efficiency in various applications. Current research emphasizes the development and analysis of multi-timescale algorithms, such as two-timescale gradient descent ascent and multi-scale transformers, to handle complex interactions between different temporal scales within data, particularly in machine learning and signal processing contexts. These advancements have significant implications for diverse fields, including reinforcement learning, financial modeling, and the analysis of complex physical systems, by enabling more accurate and efficient modeling of temporal dependencies.
Papers
Fast Two-Time-Scale Stochastic Gradient Method with Applications in Reinforcement Learning
Sihan Zeng, Thinh T. Doan
Dim Small Target Detection and Tracking: A Novel Method Based on Temporal Energy Selective Scaling and Trajectory Association
Weihua Gao, Wenlong Niu, Wenlong Lu, Pengcheng Wang, Zhaoyuan Qi, Xiaodong Peng, Zhen Yang