Yi-Fan Zhu

SAGG: Sample-Adaptive Gradient Gating for Robust Multimodal Learning under Heterogeneous Corruption

It is proved that any batch-level sample-agnostic linear estimator with a shared modulation parameter incurs an irreducible bias with respect to the clean-data gradient, and that sample-level all-or-nothing gating is the unique unbiased strategy within a natural distribution-free estimator class.