Faculty Summit 2018 – Advancing the future shape of systems toward a global AI supercomputer

已发布 2018年8月17日

分享这个页面

donald kossmann

The 19th Microsoft Research Faculty Summit yet again demonstrated its unique place in the world of computer science in gathering thought leaders, state-of-the-art ideas, new products and a sense of the possible under one roof as industry and leading academic researchers came together to share vision and purpose. More than 500 leading experts from academia, Microsoft Research and Microsoft gathered in Redmond, Washington August 1-2 to participate in the event, themed this year under the banner, Systems | Fueling future disruptions. Sessions ranged from the quest for privacy and the democratization of AI, to the future of cloud storage and the latest in quantum computing, to how to deliver AI in a secure and sustainable way, from edge to cloud to edge.

“Knowledge is the key to unlocking the potential of AI.” – Keynote Speaker David Ku, Corporate Vice President, AI Core and CTO of Microsoft AI & Research

Attendees had the opportunity to expand their knowledge via breakout sessions on intelligent edge and confidential computing in addition to AI infrastructure and tools and the latest in blockchain technology. Opportunity for dynamic interaction was real as evidenced by the buzz of hundreds of conversations in the One Table/One Topic networking lunches, with each table hosted by a Microsoft distinguished scientist and devoted to a designated research topic. Attendees also enthusiastically explored the Technology Showcase for several hours on day two of the summit and were treated to an impressive array of Microsoft innovative technologies and product ideas accompanied by quality face time with program managers and designers. As one attendee at the Technology Showcase succinctly summed up her impression of the offerings, “I’m seeing some very cool future disruptions.”

In his keynote address on morning one, Microsoft Azure Chief Technical Officer Mark Russinovich gave an in-depth talk on confidential computing and Microsoft’s devotion to the quest for a world in which customers upload and process confidential data in the cloud with the guarantee of confidential computing in the next generation of privacy-preserving cloud computing services.

Distinguished Scientist Victor Bahl, hosting the Summit’s breakout session devoted to Intelligent Edge, emphasized edge computing’s potential to complete the intelligent cloud by bringing computing near sensing and actuation. “With the introduction of edge computing, engineers can focus their creative energies towards developing a new generation of powerful, latency-sensitive AI applications that make resource constrained devices smarter, lighter and inexpensive,” said Bahl.

The buzz of very engaged attendees was perhaps nowhere more palpable than in the vicinity of the demo kiosks of the Technology Showcase, featuring everything from High Fidelity Simulations – near-realistic simulations aimed at alleviating much of the expense and danger associated with developing and testing AI in the real world – to FarmBeats, an end-to-end IoT platform for agriculture enabling seamless data collection from various sensors, cameras and drones to help boost food productivity. Throw in species recognition APIs, quantum development kits like Q#, Confidential Consortium Blockchains – and over a dozen more kiosks featuring cutting-edge product ideas and even the most determined attendees were pressed to see it all in under three hours.

“The information density is very high at this conference with lots of great projects,” said Anthony Steed, Professor in Virtual Environments and Computer Graphics at University College London as he strolled around the Technology Showcase floor. “I’ll probably spend at least a week thinking about what I’ve seen and making connections to the work that I’m doing. It’s great!”

Dawn Song, professor of computer science at UC Berkeley and CEO and co-founder of Oasis Labs, who spoke on panels including confidential computing, current trends in blockchain technology and on entrepreneurship perhaps summed up the significance of the Faculty Summit best. “It’s wonderful that Microsoft is sharing what it has been developing and what it has been thinking about and, going forward, what it wants to do and it’s great to see Microsoft putting in such an effort,” she said. “And with so many faculty here – literally hundreds — it’s great to see Microsoft enabling the interactions and collaboration between academia and industry. It’s important.”

“Our academic colleagues are the best of the best in systems and networking research and they bring a unique perspective to the hard problems we are trying to solve here at Microsoft. The knowledge and insights we shared with one another, in a free, open and honest manner, helped us grow and feel great about the future we are collectively inventing.” – Victor Bahl, Distinguished Scientist, Microsoft Research

Particularly engaging was a fireside chat on research and entrepreneurship in systems research, moderated by Ranveer Chandra of Microsoft Research and featuring Matei Zaharia, assistant professor at Stanford University and co-founder and chief technologist of Databrick, sharing the stage with Dawn Song of Oasis Labs.

In the final chat before a packed audience, Donald Kossman, Distinguished Scientist and Director of Microsoft Research Redmond Lab, delivered the perfect coda to the summit with an engaging talk on the Global AI Supercomputer and importance of a dual perspective toward understanding its significance – the internal perspective familiar to systems researchers, that of billions of heterogeneous processing and storage units and network links, but also the realization that it is one big machine from the customer point of view – and one experience — and thus the importance of systems researchers embracing this latter perspective to achieving success and trust in the marketplace.

“All our dreams are coming true,” said Kossman, referring to the advent of such a global computational platform and its significance for systems researchers across Microsoft Research and academia. “The quest now is to make the Global AI Supercomputer more secure, to create more value and to optimize and make it sustainable.”

Said Dawn Song, “I think that academia can help industry to think more about longer term research problems. And academia needs to learn from industry what are the important problems to solve. And lots of really exciting technical advancements also are happening more and more in industry today as well. This helps us to train the students to provide to industry the next generation of researchers, engineers and so on. I think this interaction is great and it’s great to see Microsoft every year putting in so much effort to bring so many faculty here.”

Next year will mark the 20th anniversary of the Faculty Research Summit. It’s certain to be a special event.

Selected Summit sessions are available for viewing on the Faculty Summit 2018 event page on the “Videos” tab.