YOU ARE AT:Analyst AngleAnalyst Angle: Why artificial intelligence is important to mobile

Analyst Angle: Why artificial intelligence is important to mobile

Artificial intelligence has progressed significantly over the past several decades, and looks set to be a boon for the mobile space

Mobile systems are relying more on cloud-based services that, themselves, are becoming more intelligent. Machine intelligence is now able to learn from experience and objectives. The next decade will be full of major benefits to society.
When I was a young graduate student in computer science at Stanford University in 1968 to 1972, I took a class from one of the most famous professors in artificial intelligence at the time, John McCarthy. Our assignments were often to write small programs in LISP, a language most noted for nesting of statements that began and ended with a parenthesis. You’d write conditions such as (IF x then y) and surround them in parenthesis and then put that condition in another one so that after a while it felt like you had 20-plus parenthesis to account for at the end of a complex task. There were even other utility programs written just to check the parenthesis to make sure they all matched up.
Our tasks were not very sophisticated in today’s terms. We’d try to figure out how to play simple games. Another professors at Stanford, Arthur Samuel, wrote a program that could play a sophisticated game of checkers. His major contribution was to build in the ability of the program to learn what moves led to a loss or a win. And, there was also a lot of effort placed on trying to figure out how to play chess – more from the position of how to figure out future moves in an efficient manner. There were a number of logic sequences that became rules for figuring out complex tasks.
Today, programming languages are much more efficient that we used eons ago. The biggest change is the use of libraries that become open and shared so when someone programs, often much of what is done is to integrate library elements to do various tasks so you can focus on solving the more difficult problem.
Here, 40-plus years later, we are beginning to see AI take the form of learning how to do something. We should really call the ability for a machine to learn “machine intelligence” rather than “artificial intelligence.” There is really nothing “artificial” about figuring out how to drive a car or fly a plane. Rather, the focus is on building “intelligence” by considering all the different conditions and then sorting through those conditions when a situation arises.
This new AI methodology is called “deep learning.” It uses a “neural network” that is able to learn as additional inputs are presented. The concept is simple, but the results can be very complex and amazing. Each node in a network has logic to look at one specific thing. It then passes the results of that analysis on to the next level that looks at something different often at a higher level. Eventually, after a number of levels are processed the goal is determined such as recognizing a face. Repeating this with new input enables the network to get better at finding the right answer.
If your car approaches an intersection, you want to have the software get inputs of the situation (what cars are present and what direction are they headed) and then make decisions that are best suited from the information. If the decision made turns out to be incorrect, then the logic is updated to make it work better for any and most every possible situation. As the logic improves, so does the intelligence. Eventually, you have millions of ways to process information through the network.
We’ll deal with the ability of software to make moral decisions later, e.g. “should the self-driving software tell the car to hit the people in the crosswalk or run the car over a cliff killing the passenger(s) in the car?”
Now, here we are in 2016, with most of the major mobile vendors – Apple, Google, Microsoft and Amazon.com to name a few – are all building software systems that are requiring them to build deep learning sitting underneath their (often voice based) interfaces to the user.
Take the work being done by Amazon for their Echo product with their intelligence engine called Alexa getting more intelligent all the time. It’s not that Alexa has to just be able to answer questions more intelligently, but the system has to also learn what you are interested in and feel is important so that it can anticipate what you might want next. In simple terms, this can take the form of realizing that you’re asking about weather alerts so, over time, the system can anticipate that you’d like to know about future weather alerts without the user asking.
A more sophisticated set of learning and anticipatory response would be to for the system to know the kind of dinners you like and then have it recommend dinners – some you like and some that are new – and if you agree, the system will order the food so that it’s ready to prepare dinner.
This same kind of logic is being researched underneath Apple’s Siri, Google’s Assistant and Microsoft’s Cortana. It’s one thing to answer a search question. That’s complex in itself, but it takes a deeper learning exercise to build a knowledge base about you and your interests and desires in order to become more valuable to you – whether it is to answer a question or offer suggestions that you might like, e.g. “I noticed that you like Brooks & Dunn. Here’s a song from Montgomery Gentry called “Gone” that you might like,” or “I noticed that you like to listen Alan Jackson. He’s coming to town next month. Shall I order you some tickets?”
Over time, our mobile devices are going to become more of an intelligence companion that not only can answer complex questions but also be proactive and offer up information, alerts, suggestions and items of interest to you as the system learns from interacting with you.
I expect that we’ll see in our lifetimes machines that can certainly drive our cars but also fly planes, help manage our finances and investments, as well as keep us fit and healthy. It’s certainly reasonable that a future Apple Watch will sense a number of things in our body and alert both us and our medical support team to help prevent us from incurring a major problem such as a heart attack.
Is it possible that machines can become more intelligence than us? I think the answer is yes for defined areas where information gathering, intelligence analysis and recommendations come into play, but we shouldn’t have to feel that our machines will replace us anytime soon. Our ability to be creative will be unique to us for a very long time.
gerry purdy
J. Gerry Purdy, Ph.D., is the principal analyst with Mobilocity LLC and a research affiliate with Frost & Sullivan. He is a nationally recognized industry authority who focuses on monitoring and analyzing emerging trends, technologies and market behavior in mobile computing and wireless data communications devices, software and services. Purdy is an “edge of network” analyst looking at devices, applications and services as well as wireless connectivity to those devices. He provides critical insights regarding mobile and wireless devices, wireless data communications and connection to the infrastructure that powers the data in wireless handheld devices. Purdy continues to be affiliated with the venture capital industry as well. He spent five years as a venture adviser for Diamondhead Ventures in Menlo Park, California, where he identified, attracted and recommended investments in emerging companies in the mobile and wireless industry. Purdy had a prior affiliation with East Peak Advisors and, subsequently, following its acquisition, with FBR Capital Markets. Purdy advises young companies that are preparing to raise venture capital, and has been a member of the program advisory board of the Consumer Electronics Association that produces CES, one of the largest trade shows in the world. He is a frequent moderator at CTIA conferences and GSM Mobile World Congress. Prior to funding Mobilocity, Purdy was chief mobility analyst with Compass Intelligence. Prior to that, he owned MobileTrax LLC and enjoyed successful stints at Frost & Sullivan and Dataquest (a division of Gartner) among other companies.
Editor’s Note: Welcome to Analyst Angle. We’ve collected a group of the industry’s leading analysts to give their outlook on the hot topics in the wireless industry.

ABOUT AUTHOR