Neural networks have been able to achieve groundbreaking accuracy at tasks conventionally considered only doable by humans. Using stochastic gradient descent, optimization in many dimensions is made possible, albeit at a relatively high computational cost. By splitting training data into batches, networks can be distributed and trained vastly more efficiently and with minimal accuracy loss. We have explored the mathematics behind efficiently implementing tensor-based batch backpropagation algorithms. A common approach to batch training is iterating over batch items individually. Explicitly using tensor operations to backpropagate allows training to be performed non-linearly, increasing computational efficiency.
AndPlus understands the communication between building level devices and mobile devices and this experience allowed them to concentrate more on the UI functions of the project. They have built a custom BACnet MS/TP communication stack for our products and are looking at branching to other communication protocols to meet our market needs. AndPlus continues to drive our product management to excellence, often suggesting more meaningful approaches to complete a task, and offering feedback on UI and Human Interface based on their knowledge from past projects.