LATEST

I’ve long been both paranoid and optimistic about the promise and potential of artificial intelligence to disrupt — well, almost everything. Last year, I was struck by how fast machine learning was developing and I was concerned that both Nokia and I had been a little slow on the uptake. What could I do to educate myself and help the company along?

As chairman of Nokia, I was fortunate to be able to worm my way onto the calendars of several of the world’s top AI researchers. But I only understood bits and pieces of what they told me, and I became frustrated when some of my discussion partners seemed more intent on showing off their own advanced understanding of the topic than truly wanting me to get a handle on “how does it really work.”

I spent some time complaining. Then I realized that as a long-time CEO and Chairman, I had fallen into the trap of being defined by my role: I had grown accustomed to having things explained to me. Instead of trying to figure out the nuts and bolts of a seemingly complicated technology, I had gotten used to someone else doing the heavy lifting.

Why not study machine learning myself and then explain what I learned to others who were struggling with the same questions? That might help them and raise the profile of machine learning in Nokia at the same time.

Going back to school

After a quick internet search, I found Andrew Ng’s courses on Coursera, an online learning platform. Andrew turned out to be a great teacher who genuinely wants people to learn. I had a lot of fun getting reacquainted with programming after a break of nearly 20 years.

Once I completed the first course on machine learning, I continued with two specialized follow-up courses on Deep Learning and another course focusing on Convolutional Neural Networks, which are most commonly applied to analyzing visual imagery. As I became more familiar with the topic, I also spent some time reading research papers and articles on machine learning architectures and algorithms not covered by Andrew’s courses. After three months and six courses, I had covered both the simple algorithms as well as many of the more complicated architectures, doing one project with each to gain a hands-on understanding.

Then I dug into the most difficult part: how to explain the essence of machine learning in the simplest possible way, but without dumbing it down. I created the presentation I wish someone had given me. (The presentation is on YouTube, where, so far, it’s been watched by nearly 45,000 people. I’ve also given it to, among others, the full Finnish cabinet, many of the commissioners of the European Union, a group of United Nations ambassadors, and 200 teenage schoolgirls to get them interested in science. Many companies have made watching my introduction to machine learning mandatory for their management.)

Thousands of Nokia employees have seen my presentation and been inspired by it. Many of our R&D folks have come to me to confess that they were a bit ashamed that their chairman was coding machine-learning systems while they had not even started. But, they said, now they were devoting their own free time to studying machine learning and were working on the first Nokia projects as well.  Music to my ears.

But that was just the first step.

The Five Steps to AI Competence

I wanted to build a way to promote a wider understanding of machine learning — not just for engineers, but for everyone at Nokia. To that end, the most valuable part of my experience has been creating a template for “The Five Steps to AI Competence.” I hope leaders across all industries can learn from these steps as they seek insights about using machine learning in their businesses:

Make everyone learn their AI ABC’s. We plan to make familiarity with the fundamentals of machine learning a mandatory process, like knowing the company’s code of conduct. We’ll create an online test. Every employee will have to study that much machine learning.

The point is not just that each individual will discover that they can understand machine learning. There’s a deeper meaning: that learning is something we need to be doing throughout our lives and that we can understand some pretty complicated stuff even if we don’t believe that at first. If we can surprise our people with their ability to learn new things, that can be very positive — for them and the company.

Create a competent pool of experts. When a business leader or anyone, for that matter, comes up with an idea — “Hey, we could save a ton of money if we did this” or “We could make this product more competitive if we could teach a machine learning system to help” — we’ll have a pool of experts to evaluate the idea and decide, “Yes, we can do it” or “Let’s try it and see” or “No way.” This could be an in-house competence center or it could even be outsourced to a third-party AI company.

These data scientists would parachute into a business unit’s generic R&D team to show them how to do what’s necessary. With every project, they’ll leave behind people who now have hands-on experience and are more knowledgeable about machine learning. They’ll spread their learning and at the same time, when they return to the centralized competence center, they can share their experience about what works on the ground.

By the way, it’s important to centralize because in today’s tight talent market, it’s much easier to recruit top talent in machine learning if they know they will be working with similarly talented colleagues.

Pair robust IT systems and data strategy. We’ll need to build IT systems that can combine any subset of data the company has access to with any other subset to amass the exact data necessary to implement a particular machine-learning system. (This may be complicated by different countries’ privacy legislation.) Setting up a data lake is pure IT work. The strategy half of the equation involves anticipating and forecasting our future data needs. In three or five years, there will be aspects of our business in which our competitiveness will be largely defined by the machine learning systems we will put in place. We’ll need to look forward to understand and acquire the data we’ll need at that time to train the systems which will be critical to our competitiveness.

Implement machine learning internally. There are numerous jobs that can be done better and faster if you augment the people working on those tasks with machine learning. For that, we’ll need to change people’s behavior so that they look at everything around them as an opportunity to automate.

Integrate machine learning into products and services. We must constantly analyze ways to leverage machine-learning to improve competitiveness with our customers.

Because these five steps are all equally important parts of the AI future, they must be implemented simultaneously. While we start teaching our employees the basics of machine learning, we can start building the IT infrastructure, searching for talent and, in partnership with our existing IT teams, work to add machine-learning competencies into our products and services. By raising the level of all the different elements of our machine-learning abilities at once, each element can connect with and enhance every part it touches. Instead of one part holding others back, everyone rolls forward together, sharing lessons, sparking new ideas and gathering momentum.

I often describe myself as an entrepreneur. When you have an entrepreneurial mindset, everything is your responsibility. You truly care and your actions communicate that loud and clear.

I could have just supported the Nokia CEO and management team in talking about the need to kick-start a fast catch-up in machine learning. But talk is cheap. Taking actions that people can see and are motivated to copy is better than any high-toned speech. The fact that the chairman of a global company went back to school and to learn a critical piece of technology was novel enough to get people’s attention and encourage them to act on their own.

I hope it’s just the beginning.

from HBR.org https://ift.tt/2PbNWu6