arXiv AI recent: Hierarchical ODE: Learning Continuous-Time Physical Prototypes for Early Link Failure Detection
Researchers proposed a hierarchical ordinary differential equation clustering network to address limitations in time series prototype learning.,The network utilizes neural ordinary differ...
The proposed method uses a hierarchical ordinary differential equation clustering network to model latent state evolution.,The network enforces temporal continuity to disentangle smooth feature trends from stochastic noise and determines the number of prototypes without rigid prior constraints.