The Evolution and Future of AI — Top Takeaways from Dr. Andrew Ng’s Visit to Adobe
Dr. Andrew Ng, cofounder of Coursera and CEO of landing.ai, and Scott Prevost, Adobe’s Vice President of Engineering, Sensei & Search
by Adobe Communications Team
posted on 08-01-2018
Recently, Adobe had the pleasure of welcoming Dr. Andrew Ng, a globally recognized leader in artificial intelligence (AI), cofounder of Coursera, and CEO of landing.ai, to Adobe for a fireside chat with Scott Prevost, Adobe’s Vice President of Engineering, Sensei & Search. During their fireside chat, Andrew and Scott hit on a variety of topics as they pertain to the Evolution & Future of AI, which ranged from education, how customers should use customer data, and the implications AI and technology have on creating a better world. Here are the top takeaways from their discussion.
There’s tons of great science going on in the field of machine learning, but Andrew expects it to mature as an engineering discipline. He likens it to how bridges used to be built—in those early days of bridge engineering, there were wise old men with accumulated tribal knowledge and reliable intuition who trained their networks to build bridges. But as bridge engineering matured, it became systematic. Andrew hopes for the same shift to happen in machine learning—that instead of being driven by tribal knowledge and gut instincts, we’ll develop principles and practices that allow machine learning to become a systematic engineering discipline.
And while the university educational systems are doing a great job teaching more and more people new technology trends, it’s equally as important that corporations do the same and help their employees with continued learning. We need a different educational system to help people to always keep on learning new things, so that someone whose job is displaced can retool themselves and learn a new trade.
One of the most fundamental ideas in the evolution of the human race is humanity’s realization that you can program a computer. It’s such a big idea and we haven’t even finished sorting out all the implications of that yet. But it’s possible that the idea that you can teach a computer would be an equally fundamental idea. Today, we have giant computer science departments focused just on programming computers. In the future, we could have equally large numbers of students and faculty focused on teaching computers.
Build unified data warehouses
One of the most important things that companies should be paying attention to about the data they collect is actually building unified data warehouses. If your data is at least put together in one database, it increases the odds that an engineer can connect the dots or run a learning app on the data to find patterns. Whereas if the data was siloed into 50 different data warehouses under the control of 50 different owners, it’s impossible for an engineer to connect the dots on the data!
Identify high value AI projects
One of the other things that companies should do is build teams that repeatedly identify the highest value AI projects and then execute against them. Although this sounds really easy in theory, the practice of how you set up the org chart, how you design the teams, how you provide the training to the engineers, can become very challenging. You have to have the technical skills but you also need to provide training to the executives. That way the executives understand enough about it to know how to advocate resources and build teams to go after the highest value projects.
Make the world a better place
In today’s world, whatever you can do in less than a second of human thought can probably be automated using a deep learning algorithm. Which means that there are plenty of things that cannot be automated, but also plenty of things that can be. The hope is that people only do work that they truly think makes the world a better place for themselves and for other people.
Every time there’s a technological disruption, it gives people an opportunity to remake large parts of the world and with this current technological disruption in AI, there is a window of time, an opportunity to change large parts of the world to be better for future generations to grow up in.
To learn about how Adobe is leveraging AI, ML and deep learning capabilities in our products, visit our Adobe Sensei site.
Topics: Digital Transformation