Spectral-Regularized Denoising AutoEncoder (SR-DAE) is proposed, designed for anomaly detection in high-dimensional spectral signals under anomaly-scarce conditions, that introduces a bandstructured spectral regularization strategy that guides the reconstruction process to preserve the coarse-grained energy distributio...
Relations in networked spatiotemporal systems are often learned from observations that entangle dynamics governed by different mechanisms, obscuring what evolves locally and how it propagates across nodes. We introduce Component-Aware Network Dynamics with Ordered Relations (CANDOR), which decomposes local dynamics bef...
Zi-Qi Wang, Dao-Jiang Hu, Cheng Bao et al.· 0 citations
Shiva-DiT, based on Residual-Based Differentiable Top-k Selection, provides four fidelity-latency operating points and reaches a 1.54x wall-clock speedup with competitive fidelity.