HP Labs Blog

HP Labs

The Future of Artificial Intelligence Research At HP Labs

By Simon Firth, HP Labs Correspondent

October 15, 2019

Tony Lewis, Vice President and Global Head of HP’s Artificial Intelligence and Emerging Compute Lab, with Palo Alto-based members of his lab.

Tony Lewis, Vice President and Global Head of HP’s Artificial Intelligence and Emerging Compute Lab, with Palo Alto-based members of his lab.

How does Artificial intelligence (AI) fit into the plans of a company that bestrides the worlds of both software and hardware? For many, “AI” references the work of building new services around search, retail sales, or social media. But thanks to its broad scope, HP can use AI to much wider effect. We checked in recently with Dr. Tony Lewis, Vice President and Global Head of HP’s Artificial Intelligence and Emerging Compute Lab, to learn more about HP Lab’s vision for using AI to build next-generation cyber-physical systems, why this is an especially exciting time to be working in the field, and how his lab is also investigating several other emerging areas of computer research.

HP: Most major tech companies are now investing heavily in AI research. What would you say is different about HP Labs’ approach?

Fundamentally, it’s that we have the ability to connect to physical systems through our printers, PCs, and other edge devices, so we can use AI in a very distinctive and powerful way. In addition to that, we’re asking how we can use AI to connect with people at a deep physiological and emotional level. We believe that this dual focus can help take HP in some new and very exciting directions.

HP: AI has been a topic of research for decades now. What has changed recently in our expectations for what it can achieve?

What’s changed is that AI is now so easy to do. Twenty years ago, if you wanted to build anything substantial, you had to do it from scratch. We also didn’t have the data that we use to teach AI algorithms in anything like the volumes that we have today. Now we have that data along with a growing set of AI tools developed by companies like Google, Facebook, and Microsoft that are making artificial intelligence networks much easier to build. That makes it a really exciting time to be doing AI research and these companies and others have already done a fantastic job of exploring the use of AI in certain well-defined domains.

But there are many more ways that we can apply AI to the world. Our specialty at HP Labs, for example, is in building physical systems, be they based around printers, other edge devices, or high-performance computers. So we’re asking: how can we enhance these physical devices with artificial intelligence to allow people to be more effective at what they want to do?

HP: Can you describe some of the specific areas where you are conducting AI research?

One is predictive analytics, where we are using AI to analyze sensor data and predict issues before they arise. We're already applying that to our printers, but it can be applied to many other research domains.

Another topic that interests us is applying AI to data that we gather about people’s physiological states in order to help them achieve their goals. We are looking at that across multiple dimensions in ways that are very different from other major research labs.

A last area that I’ll pick out it is what we call memory organization. One of the most important things we bequeath to those who come after us are records of our life experiences. But we can have tens of thousands of images on a PC or phone and no time to organize them into collections that convey the emotional power of an event. What if we could do that automatically and in a way that would help preserve our most powerful memories for future generations?

“We’re asking: how can we enhance HP’s physical devices with artificial intelligence to allow people to be more effective at what they want to do?”

Tony Lewis, Vice President and Global Head of HP’s Artificial Intelligence and Emerging Compute Lab

HP: What does answering a question like that look like in terms of the research you are conducting?

In this case it could mean helping our customers create photo books that are every bit as good as those created by professional designers. We’re already close to delivering on that vision. But we’re also thinking about a future where we will be able to preserve memories in other ways as technology advances.

HP: How far ahead are you looking as you develop your ideas?

While we do a lot of shorter term work to support our colleagues in HP’s business units, we are also tasked with looking beyond the horizon. We want to look at latent, unmet needs that we can expect customers to experience in the future and to then marry them with the technological advances that we develop. That’s both a highly technical and a highly creative enterprise – we need to use both sides of our brains if we’re going to accomplish something that will make our customers happy.

HP: Given that combination, what kind of people do you look for to work in your lab?

We’re very deliberately building a team of people with an incredible range of talents, drawn from all over the world. We have lab members in Palo Alto, San Diego, and Fort Collins in the US, and then in India, Brazil, and Bristol, England and we value the perspectives that come from people of difference ages and demographics. That includes our interns, who every year come to us with fresh perspectives that help us understand what’s most impactful in the work that we are doing. 

HP: Your lab isn’t just about AI. What other areas of research are you involved with?

The other work that we do is complementary to our AI research and falls under the umbrella of emerging compute. There we’re exploring new computing technologies and figuring out how to bring them to our customers effectively. We are working in the areas of non-volatile memory, for example, where we’re finding that we can do a huge amount to accelerate our industrial customers’ workloads when they do heavy compute tasks.

HP has also always been a company that understands the power of computing at a network’s edge. Today, edge devices like printers are doing more and more heavyweight processing. That’s moving our industry back towards a more distributed computing model, which plays right into one of our key strengths. So one of our jobs is to discover new ways of utilizing our existing assets at the edge and new ways to bring artificial intelligence to bear on the digital fabric that now underpins our world.

HP: How do you work with HP’s other research labs?

We are part of a research ecosystem where we collaborate with our sister labs to meet our shared goal of serving the needs of HP’s business groups and the global customers they work with. So we partner with the 3D lab to help them construct novel physical devices from little more than different kinds of powder. Our Print and Microfluidics lab is developing new ways to use HP’s microfluidics capability in biological applications and beyond, and we are helping them with the compute challenges they face in doing that. And it’s no good to have AI and intelligent decision making if you aren’t doing it securely. So we partner closely with our colleagues in the Bristol, UK-based Security Lab to jointly realize a vision of secure AI.

Lastly, we work closely with HP’s venture and strategy team to understand the broader scope of innovation here in Silicon Valley and our research is always informed by insights shared by our HP business partners into the global megatrends that are influencing businesses around the world and the future development of human society more broadly.