Abstract: In one embodiment, a method of simulating an operation of an artificial neural network on a binary neural network processor includes receiving a binary input vector for a layer including a probabilistic binary weight matrix and performing vector-matrix multiplication of the input vector with the probabilistic binary weight matrix, wherein the multiplication results are modified by simulated binary-neural-processing hardware noise, to generate a binary output vector, where the simulation is performed in the forward pass of a training algorithm for a neural network model for the binary-neural-processing hardware.

Abstract: A method for stabilizing forward link multiple-input and multiple-output (MIMO) rate prediction is described. The forward link channel quality is measured. Instantaneous channel quality estimates are determined for each instantaneous rank. Each instantaneous channel quality estimate provides an instantaneous rate estimate for the instantaneous rank. Rank defines the number of forward link transmission streams and rate defines the total capacity of the forward link for each rank. The instantaneous rates are filtered to determine an average rate per rank. A maximum rate is selected as the maximum over the determined average rate per rank. The instantaneous rank corresponding to the maximum rate is selected as the optimal rank. The instantaneous rate corresponding to the optimal rank is selected as the optimal rate. The optimal rank and optimal rate are then sent to an access point.