Internet of Things (IoT) has become increasingly ubiquitous during the past few years and most of these ‘things’ or IoT devices sense, measure, collect and act upon data for a specific application that they are designed for - such as ambient temperature control, vital health monitoring, environmental sensing, etc. Some of these applications deal with critical and sensitive information of a user, home, community or even a city. With deeper and deeper integration of these devices into the user’s environment, there is also a growing concern over the privacy of the data that is being collected or acted upon. Quite often, the data is also shared with other devices/things in the network, and most often sent for analysis and computation to a remote server in the cloud domain. This project aims to propose a solution that preserves the privacy of the data collected, while also allowing for data computation to be performed as close to the user or device itself instead of at the remote server every single time.
The software in this repository require two main pieces of software: Python 2.7 or higher, and TensorFlow for Python. Once they are installed, the repository can simple be cloned (using git clone) and the Python scripts can be executed.
Python: https://www.python.org/downloads/
TensorFlow: https://www.tensorflow.org/get_started/os_setup
MNIST Dataset: http://yann.lecun.com/exdb/mnist/
SVHN Dataset: http://ufldl.stanford.edu/housenumbers/