How one company is developing ‘Google for predictive analytics’

This MIT spin-out is trying to bring predictive data analytics to the masses.

Data science is a major business buzzword of late, with countless organisations examining how they can use their data to make savvy decisions.

Specifically, predictive data analytics is an area that has grabbed the attention of company leaders around the world. Using data to make accurate predictions has historically been something only large tech firms and other moneyed institutions could afford to invest in.

In recent years, a movement towards the democratisation of data analytics has begun. Siliconrepublic.com spoke to Dr Yaniv Altshuler, a researcher at MIT Media Lab, about the evolving landscape.

An idea born at MIT

Altshuler is CEO of Endor, an MIT spin-out that aims to be the ‘Google for predictive analytics’. It uses advanced AI technology based on the principles of social physics, which is a science that uses big-data analysis and mathematical laws to understand the behaviour of large groups.

Endor’s proprietary technology involves an automated engine that has the capability to swiftly analyse and predict complex human behavioural patterns. As opposed to building predictive models for each question, the social sphere of human behaviour can be analysed, enabling a democratisation of predictive science.

Social physics originates at the MIT Media Lab under the auspices of Prof Alex ‘Sandy’ Pentland, one of the world’s leading data scientists and a co-founder at Endor.

Altshuler explained: “Social physics uses big data and artificial intelligence to build a computational theory of human behaviour, whereby even complex and dynamic behavioural patterns can be captured and predicted mathematically.

“Just like physical laws governing the natural universe, social physics detects patterns in the human universe.”

User-friendly

The predictions made also stem from “fully encrypted consumer data, ensuring that clients meet privacy and security compliance regulations at all times”, Altshuler said. “Users do not even require first-hand expertise in data science in order to operate the platform successfully.”

This democratisation of predictive analytics could see the average SME gain data analytics capabilities they once thought were out of reach. At the moment, they are being left behind. “Many modern technological advancements, including predictive analytics run on big data, only benefit the largest enterprises,”Altshuler said.

“The way to reverse this trend and democratise data science is to offer affordable solutions and popularise the accessibility of data science for members of the public without data science expertise.”

Endor leverages blockchain

By integrating Endor’s platform with blockchain infrastructure, its usage aims to be cost-effective, with an eventual goal to run the platform at as low as $1 per prediction in the long term. The proprietary protocol is decentralised and trustless, having spent years in development at MIT. It will use the EDR coin as its cryptocurrency.

Endor also uses blockchain infrastructure to ensure the data used on the platform is protected. In a post-GDPR world, this is crucial for businesses seeking out ways to leverage their data.

Altshuler explained: “Blockchain technology offers solutions to this pressing issue, which is now a prominent defining factor for the reputations of multinational corporations. Handling user data safely and appropriately is a test of modernity, with many tech giants still failing.

“By leveraging blockchain infrastructure, user data can be taken out of the hands of people, who are prone to error and bias, and placed into the hands of secure, impartial technology, where all data is permanently encrypted and cannot be tampered with.”

The Endor platform itself can run predictive analytics over encrypted data “without any keys to obtain predictions, thereby preserving the privacy of the data owners and of the data subjects”.

Security is paramount

Altshuler says that data security is becoming something the general public is confronted with on a daily basis, both on social networks and any app that obtains information from them.

“Customers are more educated and aware, and public trust is fragile. Companies with reputations to protect must do extensive due diligence on any data partners they work with, and ensure their internal systems are up to the task,” he stressed.

Current users of Endor’s predictive analytics platform include large enterprises such as Coca-Cola, Walmart and Mastercard. It is opening up to SMEs in late 2018. The team has run the Endor Protocol’s beta version on big data from industries such as finance, banking, transportation, cryptocurrency transactions, gaming and telecommunications.

In the near future, the power of data analytics may no longer solely rest in the hands of major tech firms with seemingly endless budgets.

Article source: https://www.siliconrepublic.com/enterprise/endor-yaniv-altshuler-mit