Deep Learning
Social Motion Prediction with Cognitive Hierarchies
In this study, we strive to replicate this ability by addressing the social motion prediction problem.
DiffAttack: Evasion Attacks Against Diffusion-Based Adversarial Purification
Recent studies show that even advanced attacks cannot break such defenses effectively, since the purification process induces an extremely deep computational graph which poses the potential problem of vanishing/exploding gradient, high memory cost, and unbounded randomness.
Model-Based Reparameterization Policy Gradient Methods: Theory and Practical Algorithms
Based on our analysis, we propose a spectral normalization method to mitigate the exploding variance issue caused by long model unrolls.
Autodecoding Latent 3D Diffusion Models
Photorealistic generation is undergoing a period that future scholars may well compare to the enlightenment era.
Stability Guarantees for Feature Attributions with Multiplicative Smoothing
In this work, we analyze stability as a property for reliable feature attribution methods.