Ask someone their take on emerging technologies and it will revolve mostly around Machine Learning, Artificial Intelligence, Internet of Things, Blockchain etc. Certainly, these topics gathered a lot of buzz in last years. However, the technology industry is yet to take up many new innovations. One of them is “Differential Privacy”.
Technology never stops challenging itself. Large-scale data breach received much attention last year and is expected to grow. This year, we share the concept introduced in June 2016, but still not much explored – Differential Privacy. Apple introduced it at their Worldwide Developer Conference held in June 2016. Craig Federighi, senior vice president of software engineering shared at the event, “Differential privacy is a research topic in the areas of statistics and data analytics that uses hashing, subsampling and noise injection to enable crowdsourced learning while keeping the data of individual users completely private. Apple has been doing some super-important work in this area to enable differential privacy to be deployed at scale.”
Statisticians will be at the core of deploying this technology in enterprises. It encrypts personal data without disturbing the program to extract accurate data from a dataset. Simply putting, this technology would secure data from hackers to access personal information from a database. However, it will ensure that big data tools are still able to predict insights.
Apple boasts of not keeping its user’s personal information however the company completely understands the importance of this data in the era of big data analytics and machine learning. To answer this paradox, they came up with this concept of Differential Privacy. Craig also emphasized that Apple does not assemble user’s data. Apple tries technological advancements to keep the data in user’s device than on Apple’s server.
Though it is in nascent stage or we can say this is a concept, Apple is betting big on it.
This statistical science will presumably learn patterns and information about a group of people, refraining from collecting an individual’s data. Differential privacy will collect and store data in a format which extracts insights on human actions. These include their likes, wish-list, patterns, searches etc. But it cannot extract anything about a single individual. Probably we can say, that this technology will group the people with same patterns, likes and actions. It will provide useful insights but will not let attackers get access to a user’s data.
Apple announced to introduce this technology in iOS10, however researchers are debating on its success. They say that Apple has not been completely successful in implying its promise of privacy. Today, Apple is clearly the privacy leader among technology companies. So, it will be interesting to see how the tech giant will successfully implement this novel technology of data science.