Blockchain as a Service: An Autonomous, Privacy Preserving, Decentralized Architecture for Deep Learning Summary

https://arxiv.org/pdf/1807.02515v1.pdf

In today’s world of online marketing and globalized technology, many companies such as Netflix continue to use traditional software connectors and shared data storage systems in order to help serve their customers in the long run.

While the intentions of these companies for using such methods may seem pretty innocuous, the problems with these methods are that because they are expensive and time-consuming, many consumers distrust some of these companies for collecting their data. As such, companies such as Netflix cannot provide to their clients the services that they deserve.

In response to this issue, a group of four scientists: Drs. Gihan J. Mendis, Moein Sabounchi, Jin Wei, and Rigoberto Roche, decided to team up and create a decentralized data collection system that utilizes Etherium blockchain technology(i.e. Smart contracts) and deep learning methods(i.e. deep convolutional neural networks or DCNNs). This way, companies can still collect valid data from their clients and ensure that their privacy will not fall into the wrong hands, so to speak.

While this system is still undergoing experimentation, so far, the results have proven to be very successful as demonstrated by this experimentation of implementing distributed learning architectures into existing deep learning models that share data with a central controlling agent.

As of now, here in the United States companies such as Menlo One are utilizing blockchain technology in order to create decentralized apps called dApps, which are faster and more efficient than traditional computer apps.

https://www.forbes.com/sites/andrewrossow/2018/07/10/top-10-new-blockchain-companies-to-watch-for-in-2018/#678a64115600