Where is machine learning heading in 2016?

Smarter mobile apps, intelligent digital assistants and democratisation of artificial intelligence are what we can expect next year when it comes to machine learning, according to analysts.

Mark Koh from Frost Sullivan said machine learning next year will become more prevalent in mobile applications and move beyond the big guns of driverless cars, robotics and drones.

“Enterprise apps, productivity apps, and mobile apps will become increasingly smart by embedding learning capabilities or other forms of AI into the apps,” he told CIO. “For example, increasing use of NLP [natural language processing], machine vision and voice interface.”

Ian Bertram from Gartner said more digital personal assistants underpinned by machine learning will come onto the scene next year, and we are going to start to see more industry-specific digital assistants that learn over time what the user needs and the domain he/she works in.

Democratisation of AI/machine learning will continue in 2016, according to Koh. The demand to make standard machine learning algorithms accessible will see more vendors producing easy-to-use tools.

However, the skills needed to truly exploit machine learning capabilities and create more sophisticated AI will be a challenge for many organisations, with it making only gradual steps into mainstream in Australia.

“The challenge of using ML, predictive analytics and AI technologies in an automated manner is that in many cases it throws up a lot of false positives and the ability to understand the underpinning algorithm would be important – expertise that is lacking currently – to understand the results.

“Expertise in the area is going to hinder adoption for more sophisticated use of AI and the big players will continue to be the ones driving this development for specific applications,” said Koh.

Bertram pointed out that buying an off-the-shelf machine learning tool or predictive API doesn't replace the skills needed to continually tune the algorithms, shape the data and program a machine learning model to properly work in production.

“The skills to do machine learning, the data science skills to do machine learning, we are still in a short supply globally,” he said.

“I can’t just go out and buy a machine learning thing and plug it into my system, and all of a sudden I’ve got this great environment. There’s got to be someone within the organisation, or a group within the organisation, that knows how to program this thing, has the skills to develop this thing and continually adjust it, tune it.”

Stuart Johnston from Deloitte said this won’t stop many companies from thinking about how they could use machine learning to drive their business forward in the coming year, and start adopting some of the tools and technologies.

“Machine learning in the form of 'how can organisations use algorithms more?' – executives have to be thinking about that," Johnston said.

"It really is the next wave of digital disruption. It will impact and change business models, and if executives – especially our top companies – are not thinking about machine-based learning, then they run the risk of their businesses being disrupted in the next three to five years. It’s much in the same way as digital disrupted businesses in the last decade.”

More companies are also going to go beyond simply measuring data and collecting metrics to more algorithms underpinning some of their operational and customer-facing technologies, said Kristian Steenstrup from Gartner.

“Those Fitbit wrist bands, for example, that have been around for a couple of years now… use quite advanced proprietary algorithms to interpret the data. It’s not just measuring things, it’s using algorithms to interpret it.

“These sorts of things can work out when you are going up vertically, and it knows the difference between drinking beer at a bar and swinging your arms when you are walking. It can because of algorithms interpreting the rate of speed, distance and movement.

“If you extrapolate that out to other areas, then companies that have really good algorithms to interpret data will have an advantage. And I think we’ll see proprietary algorithms as a differentiator in business,” concluded Steenstrup.

Check other trends for next year:


Copyright © 2015 IDG Communications, Inc.

6 digital transformation success stories