Where Machine Learning is Headed in 2018 And How India’s Tech Sector Plans To Keep Up ?

After decades of false dawns, in the past couple of years the excitement around artificial intelligence seems to have reached something of a fever pitch. Aided by rapidly increasing computing capabilities and an unprecedented volume of real-time data, computer scientists have all the necessary elements in place to unlock the age-old potential of neural networks for the first time ever. By setting these factors to work, they’ve managed to train machines to recognize valuable information and even use it to draw actionable insights that actually impact our day-to-day lives.

Almost overnight, speech and image recognition have become far easier, to the point that virtual assistants like Apple’s Siri and Amazon’s Echo are commonplace. These advances have only become more pronounced with the introduction of smart sensor technologies that can translate real-world information into digital data. Now suddenly, self-driving vehicles and automated workforces have become a very real possibility.

Of course the progress isn’t showing any signs of letting up, many experts in the space have pegged 2018 to be the year where many of these machine learning breakthroughs actually come into play. Here are some of the most hyped trends in the coming year.

Machine Learning in Businesses

According to Gartner 59% of organizations are currently in the process of adopting a concerted AI strategy. While businesses have already begun experimenting with these technologies through social media listening tools, chatbots and smart search engines; 2018 promises to bring machine learning to the enterprise in a big way.

With Google and Amazon set to release self-service machine learning portals that allow businesses to train algorithms based on their own specific data. Armed with these capabilities, organizations will look to create automated systems that can bring together intelligence from all functions of the enterprise and use these insights to drive optimization in everything from logistics to HR.

Increasingly, organizations will look to bring together different forms of data to create combined insights. For example real-time information from commodities markets could be compared against satellite imagery to create a holistic trading model which allows companies to stay one step ahead of the market.

Capsule Networks

While traditional neural networks have been able to build complex images of individual objects for a while now, they have struggled to identify these objects when they are moved to a different location or to place them in a larger spatial context. Capsule networks can identify the same patterns as traditional networks with far more limited information. Now, instead of exhaustively training a machine to spot a single object at all possible orientations, data scientists should be able teach these machines to recognize objects in context.

India's Response

With the recent Digital India initiative taking off with full force, India’s IT sector is set to grow by at least 50% over the next couple of years, with at least 180,000 more jobs expected in the current year alone. Based on Gartner’s survey of Indian CIO’s leading companies are willing to commit more than a third of their budgets towards digital transformations involving analytics and cloud infrastructures.

This shift in focus is translating directly into a growing demand for machine learning specialists, with data scientists in particular leading the pack. According to Kamal Karanth a Co-founder at the recruitment company Xpheno, open vacancies in data science currently number around 50,000, and this number is expected to double over the course of the year.