What Is Machine Learning? Your Smartphone Knows
In the near future, machine learning chips will make smartphones even smarter, allowing your device to “act like a human” and anticipate future outcomes based on past events. No more big operating system updates – soon, your phone will diagnose its own needs and improve itself.
“AI promises to be the most disruptive class of technologies during the next 10 years due to advances in computational power,” said John-David Lovelock, VP of Research at Gartner, Inc. In 2018, AI technology grew at an estimated rate of 70 percent. And it will keep growing: by 2022, the global value derived from AI is projected to reach $3.9 trillion.1
Artificial intelligence will likely be standard across smartphones by 2023. In the meantime, read on to get ahead of the curve.
Basics of Machine Learning
Machine learning is a type of artificial intelligence (AI) that allows computing systems to progressively improve performance based on data, without being programmed to do so. Humans are built to progressively improve our performance based on past experience (we just call it “learning”). Soon, personal devices will be, too.
While human learning relies on inference and judgment, machine learning algorithms make predictions solely based on data. For instance, it takes many, many text inputs for a word processing or text messaging app to reliably correct typos and predict words.
Machine Learning in Smartphones
Your smartphone already contains sophisticated sensors and the ability to analyze data from those sensors. It can lower its screen brightness in response to ambient lighting, check its own battery life and turn off the display when not in use. But machine learning will allow it to predict future behavior based on past usage, adapting in real time as it “learns” from its experiences.
Dedicated artificial intelligence chips will be at the core of smartphone machine learning. These chips will not only analyze data, but form a relationship with the device. The on-device machine learning (as opposed to cloud-based machine learning) made possible by these chips will preserve privacy by limiting the data sharing required to “train” an AI program.
With a dedicated chip, all improvements will happen natively and internally – no cloud, no wireless and no need to download OS updates.
1Gartner Newsroom, 2018. “Gartner Says Global Artificial Intelligence Business Value to Reach $1.2 Billion in 2018.” Available at https://www.gartner.com/newsroom/id/3872933.