We lift the theory of optimal reduction to a decomposition of the lambda calculus known as the Linear Substitution Calculus (LSC). LSC decomposes β-reduction into finer steps that manipulate substitutions in two distinctive ways: it uses context rules that allow substitutions to act “at a distance” and rewrites modulo a set of equations that allow substitutions to “float” in a term. We propose ...

Traditional convolutional neural networks (CNN) are stationary and feedforward. They neither change their parameters during evaluation nor use feedback from higher to lower layers. Real brains, however, do. So does our Deep Attention Selective Network (dasNet) architecture. DasNet’s feedback structure can dynamically alter its convolutional filter sensitivities during classification. It harness...

Large ZnSe nanocrystals are expected to be promising blue-light emitters with an emission peak of 455–475 nm, which is important for the construction display apparatus. The final size via one-step injection can varied by reactivity Zn and Se precursors; however, it has a limit <5 nm. To describe key factors in determining nanocrystals, we proposed nuclei number-considered LaMer model based on M...