Motion-based AI video detectors exploit motion biases in evaluation datasets and drop to near-random performance on rebalanced data, while frequency-based detectors remain robust.
Perceptual straightening of natural videos.Nature neuroscience, 22(6):984–991
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
LeWM is a ~15M-parameter JEPA world model that trains end-to-end from pixels with only next-embedding prediction plus a Gaussian latent regularizer, cutting loss hyperparameters to one.
citing papers explorer
-
Dataset Biases and Shortcut Learning in Motion-Based AI-Generated Video Detection
Motion-based AI video detectors exploit motion biases in evaluation datasets and drop to near-random performance on rebalanced data, while frequency-based detectors remain robust.
-
LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels
LeWM is a ~15M-parameter JEPA world model that trains end-to-end from pixels with only next-embedding prediction plus a Gaussian latent regularizer, cutting loss hyperparameters to one.