Tag
The paper proposes Mass-Aware Attention (MAA), a generalization of standard attention normalization that retains evidence accumulation information by scaling numerator and denominator at different rates under repetition, improving representation informativeness and task performance across temporal graph and other domains.
Introduces Neural Bayesian Sequential Routing (NBSR), a framework that models neural inference as sequential evidence accumulation over a DAG using Dirichlet-Categorical conjugate updates, enabling uncertainty quantification, early exiting, and resource-rational inference.