von Mises-Fisher

A new arXiv paper (2609.30647v1) formalizes visual representation learning as a conditional predictive sufficient statistic (CPSS), showing that predicting the next patch embedding with a cosine loss …

Researchers propose Gradient Mirage, a defense against gradient matching attacks (GMAs) in large language model (LLM) split learning, which breaks the gradient-objective consistency that attackers exp…