Menu
Menu
Will machine learning become part of our everyday lives?

Will machine learning become part of our everyday lives?

Carlos Guestrin, who’s been practising machine learning long before it was cool, expects every application to have some form of embedded intelligence in five years

Machine learning, schmine learning – it’s just going to be a part of our everyday lives. That’s what Amazon professor of machine learning at University of Washington and Dato founder, Carlos Guestrin, says about this artificial intelligence subfield.

Guestrin has been practising machine learning long before it was cool, having started out as a student in robotics and automation engineering at Universidade de São Paulo in Brazil in the early '90s, before moving to US to attend Stanford.

He worked for Intel and as a professor at Carnegie Mellon University and University of Washington before starting his own company, Dato, which provides machine learning software and training.

"In the next five years, every successful breakthrough app is going to use machine learning at its core. Machine learning is what’s going to make an app truly useful and different to other things out there,” he said in the lead up to the Strata conference in New York City.

"And users are actually expecting this. It’s not just that the technology has evolved and the amount of data collected has increased, but the expectations of consumers have changed over the last several years.”

Guestrin says machine learning is already making its way into a wide array of industries such as healthcare with personalised medicine, to behavioural targeting in marketing, to smart warehousing and smart urban planning. Whether it’s booking a ride, or finding the next song or movie, machine learning can be embedded in just about every single app, he said.

“It used to be the case we’d talk about a phone as just being a device, but now we think about a smartphone and almost personalise it. If it doesn’t react in a personalised way to me, I get upset about it. I really want it to understand my needs, I really want it to provide me with an individualised experience. So that’s really only possible with machine learning.”

One area where machine learning will become indispensable, said Guestrin, is with Internet of Things applications, such as those in the home that schedule and monitor appliances.

“I have worked in this area for many years. Early in my academic career it was called ‘sensor networks’ or ‘network of wireless sensors’. With a home automation system, we would want it to predict our needs or react ahead of time to our needs and what our interests are and what the current situation is.

“The only way to do that is to gather the data, automate the adaptation process and continuously adapt to how things are evolving,” he said.

Respect for privacy, however, must be top priority if machine learning is to be embedded in almost every app, Guestrin stressed. The convoluted, drawn out terms and conditions are on the way out; users need clear and transparent statements as to how their data is going to be used by companies.

“If you help people understand when machine learning is being used and what is it being used for and the value it provides to [the consumer], then it becomes more clear to the users what is happening under the hood. By making it more clear, we can make this process much more effective where it’s not creepy. So it all becomes more honest and open.”

Making machine learning accessible

One barrier to widespread industry adoption of machine learning is skills and resources. Most companies are not the likes of Google that can hire large teams to build innovative, new intelligence applications, Guestrin said.

Therefore, we can expect to see an increase in machine learning tools and APIs to address this need, he said.

By making machine learning more accessible to businesses and to the broader community, it will help spur on new innovative apps and allow people to focus on creation rather than it only being an academic discipline.

“One analogy I like to make is the iPhone/iOS app store, which made it easy for a wide range of people to create interesting phone applications using the underlying iOS API. If we can make machine learning tools accessible to a wider community, then you can only expect creativity and let more folks build really cool and interesting things from it,” Guestrin said.

Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO newsletter!

Error: Please check your email address.

Tags Carnegie Mellon Universitymobile applicationsUniversity of Washingtonpredictive modellingCarlos Guestrin. machine learning

More about GoogleIntelMellonStrata

Show Comments

Market Place

Computerworld
ARN
Techworld
CMO