/Richard Uhlig Talks 2021 Tech Trends (via Qpute.com)
Intel silicon photonics future technology

Richard Uhlig Talks 2021 Tech Trends (via Qpute.com)


If you imagine Intel as a giant restaurant chain, serving people all over the world with not only great food but also much needed fuel and sustenance, it wouldn’t be hard to picture Richard Uhlig as a masterchef tasked with discovering culinary breakthroughs and developing new recipes with a team of exceptionally talented chefs.

Intel chip
Getty Images

But picturing him as a chef restricted to his kitchen alone or a scientist confined to Intel’s super-secretive research base would be a mistake. Eureka moments don’t just happen in a lab, ideas don’t always follow a predicted path, and Uhlig knows this well. Sometimes inspiration strikes beyond the tried and tested, among fresh minds with newer perspectives.

“We have a budget that we apply on a rolling basis to fund research in universities across the planet, encouraging investigations in different themed areas that’s important and strategic to Intel,” Richard tells me.

“We don’t just fund a research grant, but we actually design research centers under given themes. And we also have our own researchers sometimes even physically present at the universities, doing joint research with the academics, often publishing papers together,” emphasizes Richard Uhlig, while explaining how Intel’s academic research grants have a three-to-five year period with laser-focused intent.

Academic investments are really important for helping Intel decide what’s the next big thing to focus on, Richard says, “Because that’s where we get all the signals on what are the key trends and it also plays a big role in bringing fresh talent into the company. Those academics that we fund have grad students that invariably become future Intel Labs employees.”

Finding key ingredients

As an Intel senior fellow, vice president and director of Intel Labs, Rich Uhlig is tasked with “exploring the future for Intel” with little to no strings attached. Richard’s team is unique because it goes outside Intel to not only engage with academic researchers but also build meaningful partnerships with government agencies, research organizations and industry partners.

From an organizational point of view, what Intel Labs is trying to do is make sure they have the right composition of diverse talent in every respect to lead the charge of defining and developing Intel’s future technologies roadmap, according to Richard Uhlig.

“We want diversity of different geographies, different technical domains, different backgrounds, and overall just researchers with different life experiences,” says Richard, before continuing. “I really believe that if you want to innovate, you have to work at the intersection of domains. And oftentimes that’s where innovation happens.”

Intel Labs Richard Uhlig
Dr Richard Uhlig, Director, Intel Labs (credit: Intel)

Richard Uhlig’s unique mandate at Intel Labs means he doesn’t only set his own agenda but also plays a crucial role in defining Intel’s future technology portfolio.

“We essentially cover everything from the transistors up, so we look at microarchitecture to system architecture, from system software to programming systems, we examine matters related to security and design efficiency of our products and solutions, and we contribute to Intel’s future technologies in each of those domains,” says Richard, pointing out how Intel Labs serves the company as technical advisors.

“Based on the things we learn from our external engagements and our own internal research, we make recommendations to Intel’s top leadership in terms of technology trends the company should be paying attention to or developing in the future,” he says.

Refining the palate

One can only guess the number of future technologies and breakthroughs Intel’s currently baking in the oven, but Richard does point to a chosen few that’s almost ready to be put on the menu for all to enjoy.

Cutting-edge research in domains like integrated photonics, neuromorphic computing, quantum computing, confidential computing, and machine programming is especially exciting, according to Richard Uhlig. Because all of these areas present critical challenges in the future of computing, which require pioneering efforts and real-world breakthroughs to overcome, something that’s right up Intel’s alley.

If you take the example of integrated photonics, it presents a fundamental problem that Intel’s working on for over two decades now, something that’s increasingly manifesting in data centers all over the world. The challenge of reducing energy consumption in high-bandwidth, high-performance compute workloads.

“The power consumed while delivering data to high performance compute engines is consuming much of the power budget in data centers,” explains Richard. The answer to this problem is using light or photons to carry information instead of electricity where it matters the most.

Intel silicon photonics future technology
Intel innovating in the field of silicon photonics (credit: Intel)

After several years of core physics research and development of different components required to implement photonics into existing compute nodes, “Intel has began to ship silicon photonics transceivers as discrete parts that are being actively used by many cloud service providers for long-range server rack communication within the data center,” Richard says, claiming how this is very much the need of the hour to not only increase data transmission rates through photons but also save on spiraling energy costs.

“We have an opportunity to do another round of innovation through silicon photonics, where we’re looking at ways of integrating photonics directly into a compute package,” Richard mentions, pointing to the next stage of even tighter integration of silicon photonics technology right down to the chip-level of future Intel products, in the quest to deliver on an ambitious goal initiated many years ago to tap light as the basis of connectivity technology.

Beyond silicon photonics, Richard highlights how Intel’s also pursuing general purpose compute models that go beyond the traditional x86 architecture, by developing entirely new ways of thinking about how computers and computing will evolve.

Developing a cuisine

“That’s exactly what Intel’s neuromorphic and quantum computing programs are all about, and they’re each in their own way exploring different models of computation that we think have unique properties and use cases,” says Richard.

Intel Loihi Neuromorphic Computing Chip
Intel is advancing neuromorphic computing (credit: Intel)

“In the case of neuromorphic, we’re inspired by the energy efficiency of biological brains and their ability to learn with relatively smaller number of examples in contrast to training an artificial neural network in deep learning scenarios,” he says, “while quantum computing is all about controlling some of the intractable problems related to qubits and other fundamental quantum building blocks.”

Intel’s Neuromorphic Research Community has already shown some very interesting results in the areas of artificially mimicking the human olfactory system and bringing event-based touch sensing to robotics, strengthening the case for neuromorphic computing being well-suited for an emerging class of bio-inspired intelligent workloads that also have commercial relevance through Intel’s neuromorphic research test chip, Loihi.

In the quantum realm, Intel’s second-gen Horse Ridge Cryogenic Quantum Control Chip builds on the first-generation Horse Ridge controller introduced in 2019, as Horse Ridge II supports enhanced capabilities to manipulate and read qubit states and entangle multiple qubits. All of this slowly builds momentum into much bigger, more disruptive quantum computing gains just around the corner, Richard highlights.

Intel horse ridge 2 quantum control chip
Intel horse ridge 2 quantum control chip (credit: Intel)

“Beyond this, Intel Labs is also pursuing advancements in confidential computing, with federated learning and homomorphic encryption technology, which will allow us to learn from data segregated by silos without impacting its integrity,” Richard explains.

“We’re also trying to solve the increasing challenge of programming complex computing systems in a world where highly skilled programmers are becoming a scarcity, with something called machine programming — where we use AI methods to teach computers to program themselves. This is a relatively new research theme for us, but one that we’re quite excited about to make some important contributions,” he says.

When life gives you lemons

COVID-19 obviously derailed the growth engine of the world, not only locking down markets and economies but also throwing a massive wrench into the roadmaps of companies like Intel. Or so you’d like to think.

“When COVID-19 started to take over the world, we just took a look at the work that we were already doing in the labs to see if it can help us adjust to the new normal,” Richard describes, while speaking about technologies like remote telepresence that create more immersive digital interactions, something that can benefit the entire work-from-home phenomenon that’s finding more acceptance due to COVID-19.

“Imagine a computer display that has multiple cameras right around its bezel, which can take multiple perspectives of a person sitting in front and integrate them to reconstruct a more realistic image on the viewer’s side when it comes to tracking tiny, more believable movements. It’s like you’re looking at someone through a window rather than a flat screen. This is one example of something we were working on that we weren’t really sure exactly where it might go, until COVID-19 happened,” Richard reveals.

Intel machine programming
Intel unveiled ControlFlag – a machine programming research system that can autonomously detect errors in code

The more Richard Uhlig speaks about the scope of his work at Intel Labs, I start picturing him as someone trying to solve a ten-by-ten Rubik’s cube in record time or putting together a thousand-piece jigsaw puzzle, asking himself how silicon photonics fits into quantum computing or neuromorphic computing employs federated learning, among other things. He smiles before answering, “We’re always looking for those synergies across domains and that goes to what I said at the beginning about bringing together people across different areas so that they can figure out how the pieces of the jigsaw puzzle come together.”

“At the same time, I think it’s important not to force fit things, you know? We want to make sure that we have a broad, diverse portfolio of ongoing investigations. And we look for those opportunities for things to come together when these technologies reach a tipping point; but we don’t force it, if it’s not natural,” Richard emphasizes.

The recipe for success

After earning his Ph.D. in computer science and engineering from the University of Michigan back in 1995, Dr Richard Uhlig joined Intel in 1996 and is currently in his 25th year at the company. Even after so much time, Richard’s never been more confident about Intel’s innovation roadmap and its ability to keep pushing the technology needle forward.

“Intel is a technology company and we have an amazing population of technologists across a very broad range of disciplines that’s aspiring to create opportunities and go after newer markets and that’s the core of what’s going to help us succeed into the future,” he says.

Richard Uhlig of Intel Labs
Dr Richard Uhlig (credit: Intel)

“As the director of Intel Labs, my role personally is to create the environment where this innovation can happen, to have the right dynamic in the organization so that the brilliance that’s present is able to manifest itself and truly flourish,” Richard sums up, in no uncertain terms.

Not a masterchef, nor a complicated jigsaw puzzle solver, Richard Uhlig sees himself as a gardener, “making sure the garden and all the plants that are growing in it are healthy, well fed and nourished, and ultimately engineered for success.”

.


This is a syndicated post. Read the original post at Source link .