Download Research Tools
The first week of July didn’t just see the arrival of extraordinarily high temperatures across Europe—it also brought extraordinarily high energy to Microsoft Research Cambridge, UK, as 81 top PhD students gathered for the tenth annual Microsoft Research Cambridge PhD Summer School. Hailing from 35 research institutions spanning 16 countries in Europe, the Middle East and Africa, the students brought huge diversity to the Cambridge Lab, not just in terms of national origin and culture, but also in their research backgrounds, which extended beyond computer science and engineering into the realms of design and various natural and social sciences. The event’s attendees included recipients of Microsoft Research PhD Scholarships, along with students whose work involves our EMEA Joint Research Centres and those who are collaborating on Microsoft Azure for Research projects or are otherwise partnering with Microsoft Research Cambridge.
Microsoft Research Cambridge Laboratory Director, Andrew Blake, opened the Summer School, welcoming the students before they launched into an ambitious four-day agenda—a carefully designed mix of scientific talks and demonstrations, training sessions and other practical activities, and social events that offered lots of opportunities for networking.
Following are the highlights of this year’s PhD Summer School.
Talks from invited experts
The invited talks began with considerable excitement when Hermann Hauser, a serial entrepreneur and co-founder of such high-tech companies as Acorn Computers (famous for the BBC Micro, which dominated the UK educational computer market in the 1980s, and later spun out ARM and a number of other companies), gave a talk entitled “Technology Development.” Hauser first described earlier waves of computing, before concentrating on machine learning and artificial intelligence as technologies that could transform not just our economy but every aspect of our future lives.
Other invited speakers included Marta Kwiatkowska, professor of computing systems at the University of Oxford, who explained aspects of DNA computing in her talk entitled “Computing Reliably with Molecular Walkers,” and Bernhard Schölkopf, director of the Max Planck Institute for Intelligent Systems, who delivered a talk on “Empirical Inference in Intelligent Systems.”
The researcher talks and demonstrations gave a broad overview of the activity in the Cambridge Lab, from environmental science and computational biology, to various aspects of core computer science, design and human-computer interaction. The students were especially attentive—you might even say awestruck—when Microsoft researcher and Turing Award winner Sir Tony Hoare spoke on “The Laws of Programming with Concurrency,” giving a historic account of his work on Hoare logic and communicating sequential processes (CSP).
Researcher talks gave the students an overview of the cutting-edge work underway at Microsoft Research Cambridge.
A special highlight was the Code Hunt contest, during which the attendees could demonstrate their C# or Java programming ability by solving a series of increasingly complex coding puzzles. About half the students participated, and the winners received special acknowledgement at the formal dinner on Thursday evening at Jesus College Cambridge.
The training activities are an all-time favourite in the Summer School agenda. Besides professional coaching on how to deliver a research presentation and how to present a poster at an international conference, students heard lectures from senior Microsoft researchers, who shared their general learnings on research and how to pursue a successful research career.
Lunchtimes provided food for the brain as well as the body, as the PhD students presented their posters to the 100-plus researchers (and almost as many summer interns) of the Cambridge Lab. Presenters got invaluable feedback from their peers and the Microsoft researchers, and also benefited from the insightful advice of poster coach Sue Duraikan from Duraikan Training, a consultancy that provides support in designing and delivering learning strategies.
Even PhD students need to eat, and they dined in style at the formal dinner at Jesus College Cambridge. Ah, and you thought this was a photo from Hogwarts!
The 2015 PhD Summer School closed with a packed lunch on Friday—not every meal can be as refined as the dinner at Jesus College! As in other years, I was sorry to see the event end, since I really enjoy getting caught up in the excitement of these students who are on the cusp of great careers. Fortunately, I can look forward to next year’s event, and hope that the weather will once again put the “summer” in the Microsoft Research Cambridge Summer School.
—Scarlet Schwiderski-Grosche, Senior Research Program Manager, Microsoft Research
The 2015 Microsoft Research Faculty Summit is over, but I am still recovering my voice from all the great hallway conversations! The summit reminded my fellow Microsoft researchers, our myriad collaborators in academia, and me of what we have already accomplished and the exciting opportunities ahead.
Harold Javid blogged about day one of the summit in detail, but I would like to call out a few of my own takeaways from that day. First, I was extremely pleased by the reactions to Jeannette Wing’s announcements of the RFPs for Microsoft HoloLens and Catapult. I overheard my academic colleagues discussing ideas for taking advantage of these grant opportunities. I enjoyed the spirited discussions that characterized Eric Horvitz’s panel on artificial intelligence—there’s a palpable excitement about the entire AI field that Eric’s panel captured. You can watch on demand Jeannette’s keynote and Eric’s session.
The Universal Design breakout session offered some fresh motivation and approaches for diversity, which Harold covered in his blog. Although diversity is my passion, Professor Charles Isbell of Georgia Tech delivered an “aha moment” for me. More than 60 percent of the faculty at the top 4, 10, 20, and 25 computer science programs are graduates of one of these same programs. The ranking of one’s PhD institution is a huge factor in hiring—departments hire at their own rank or higher. This is common knowledge, but Charles connected it to diversity. If the very top programs would make a truly concerted effort to increase the participation of women and minorities in PhD programs, the effect would propagate throughout the entire computer science field. Only a few people, those who lead and serve on the PhD admissions committees, can make it happen.
Finally, I really enjoyed socializing that evening with colleagues from around the world. It was great to see so many old friends—and to make new ones. I was especially excited that so many assistant professors attended for the first time and were enjoying the research breadth of the content and making new connections. My abiding memory of the reception will be the way Harry Shum used his “rock star” status to spark animated conversations throughout the night.
On day two, I had the honor of introducing one of the most influential researchers in my field, Stanford professor Monica Lam. Monica delivered a provocative keynote on Omlet, an app that she and her students at Stanford created as an open alternative to Facebook and China’s WeChat, which she indicts as “big-brother” social networks that trample privacy and exploit users’ personal data to their own profitable advantage. I could try to summarize her arguments, but I wouldn’t do them justice, so I encourage you to listen to her presentation for yourself. Everyone was hungry for more technical details.
Monica Lam delivered her keynote, A Revolution Against Big-Brother Social Networks.
After Monica’s address, it was time for the Research Showcase. This year, we expanded the showcase to include 47 demo and poster booths with a range of exciting projects underway at Microsoft, many of them joint efforts with academic institutions. There was truly something to interest researchers from every domain of computer science. The demos were clustered around six broad themes—Artificial Intelligence, Software, Devices, Computing in Society, Research in Action and Engaging With Microsoft—and featured crowd-pleasers like the RoomAlive Toolkit and cutting-edge topics like What Can We Solve with a Quantum Computer?
My own favorite was the demo of an AI project that can accurately guess your weight, waist size and body mass index (BMI) based on a handful of inputs. (The estimate was over in my case!) The TouchDevelop demo showed some of the software capabilities of the tiny Microbit programmable device that will soon be in the hands of every middle-school-aged student in the UK. The booth on Programmability of Scalable, Geo-Distributed, Interactive Applications—surely in the running for the wordiest name—is using a variant of the actor model, which was invented by Professor Gul Agha of the University of Illinois at Urbana-Champaign, who enjoyed the demo with me and had some great feedback for the interns! Serendipitous moments like this are what make the Faculty Summit special.
The Research Showcase featured 47 demo and poster booths with a range of exciting projects underway at Microsoft, many of them joint efforts with academic institutions.
More than 400 summit attendees viewed the showcase, and a number of them voted for their favorite demos. The results of this highly unscientific popularity contest were as follows:
After the showcase and lunch, it was time for more breakout sessions, including one I led on Programming Models for Estimates, Approximation, and Probabilistic Reasoning (maybe I shouldn’t have been so hard on that wordy demo title). A panel of experts—Noah Goodman of Stanford University, Dan Grossman of the University of Washington, Michael Carbin of Microsoft Research and MIT, and Todd Mytkowicz of Microsoft Research—provided multiple perspectives on how to deal with the problems that result from using data from sensors, machine learning, approximation and crowdsourcing, all of which have errors. From programming to type theory to probabilistic reasoning, there were enough equations to satisfy even hardcore mathletes. Their presentations attracted a very diverse set of researchers, including experts and theorists in programming languages and AI, and generated a lot of pointed questions.
Then I was off to sit in on the biggest draw of all the breakout sessions—the standing-room-only discussion of The Holoscene: Virtual and Mixed Reality. This informative and entertaining presentation traced the history and promise of holography, and talked honestly about the problems, both technical and social, that remain to be solved before we all begin living in a mixed environment of the virtual and physical worlds. You will be able to view this session and many of the other breakout sessions online soon. If you’d like to review the presenters’ slides, many of them are already downloadable from the agenda page.
Peter Lee, corporate VP of Microsoft Research, delivered the concluding keynote.
The concluding keynote from Peter Lee, corporate VP of Microsoft Research, was aspirational. He described the philosophy and example research projects in our new research division at Microsoft, called NeXT (New Experiences and Technologies), which he is leading. NeXT aspires to create a new model for conducting research and technology transfer—with joy, risk taking and fearless dedication to discovery. He exhorted the audience to double-down on basic and disruptive research, especially today, when politicians and the public often demand to know the ROI for their research investments. He ended his talk with the metaphor of research as a rollercoaster ride: it’s terrifying as you climb up the first hill, but exhilarating once the ride gets rolling.
I hope all of you will heed Peter’s advice—create your own rollercoaster!
—Kathryn S. McKinley, Principal Researcher, Microsoft Research
On May 27, 2015, Suzhou, a historic, scenic city in southeast China, hosted this year’s finals of the Beauty of Programming (BoP) competition, a creative contest among talented young programmers. Sponsored by Microsoft in collaboration with IEEE, BoP encourages contestants to use their programming skills to solve practical problems.
With a theme of Artificial Intelligence (AI), BoP 2015 attracted more than 22,000 online contestants from more than 150 universities. Code Hunt, a game in which players demonstrate their coding skills by solving intriguing puzzles, served as one of two online challenges—hihoCodeer was the other—that narrowed the field to the top performers.
BoP 2015 attracted more than 22,000 online contestants from more than 150 universities.
Fifty students made it to the final round in Suzhou, where they undertook either of two challenges designed by Microsoft employees. One, called Cortana Next, challenged contestants to create a voice assistant by using APIs from Project Oxford. The other, named News Recommendation, asked the participants to sort out the most valuable news from 3,000 items.
Feihu Tang of Harbin Institute of Technology and Xiaoxu Guo from Shanghai Jiao Tong University earned first place. They chose the Cortana Next challenge and developed a travel app they named RainbowGO, which uses speech recognition, natural language processing and scenario analysis to turn Cortana into a proficient trip adviser. RainbowGO analyzes a traveler’s departure and destination sites to provide time conversions and information on round-trip tickets, nearby attractions and public transportation. Tang hopes to parlay his experience at BoP into someday starting his own business.
Feihu Tang demonstrates his app to Microsoft employees, including Yongdong Wang, CTO, Microsoft Asia R&D (blue shirt in the middle)
Xiaoqi Chen of Tsinghua University won second place for his News Recommendation app that uses the text segmentation API from the ltp-cloud.com and SVM algorithm to sort out the most valuable news. Chen and his partner, Junyan Li from Chinese University of Hong Kong, also took one of the two awards for Best Demo.
“I’m really excited by coding under a 12-hour deadline to solve a certain problem, and I enjoy the experience of pair programming,” said Chen, who was competing in his second BoP. He praised the contest as a chance to partner with an accomplished programmer.
Junyan Li (left) and his partner Xiaoqi Chen (right), who was competing in his second BoP and got second place this time.
Chen adds that he was eager to compete against peers from India, which brings up an important aspect of this year’s BoP. The competition has always sought contestants with diverse backgrounds, but this year marked the first time that it included participants from India and Japan. Four enthusiastic contestants from Japan and India won the Best Team Award, and the Indian team also took home one of two Best Demo Awards.
“This was a once-in-a-lifetime opportunity for me. The competition was very well organized and challenging. It’s such a good format and maybe it can go to the next level, like having more foreign teams in the future,” said Chirantan Mahipal, one of the winners from India.
“We hope to include more students from other countries in future BoP contests,” said Xin Ma, a senior research program manager at Microsoft Research Asia. “They are passionate about programming and add a lot to the brainstorming; after all, programming has no language barrier.”
Haoyong Zhang, a principal development manager at Microsoft’s Suzhou Center, noted that recent developments in the mobile Internet, big data and natural language processing contributed to the 2015 BoP theme of AI. “I hope contestants broadened their horizons through BoP, contemplating the revolutionary changes that ‘geek power’ and AI can make in the real world.”
Although the Beauty of Programming 2015 contest has ended, the contestants’ passion for computer programming remains. These talented young programmers are sure to have more opportunities to engage in innovative projects. We at Microsoft derive inspiration and motivation from the enthusiasm of these aspiring IT developers.
The 50 contestants who made it to the final BoP 2015 round in Suzhou
—Guobin Wu, Senior Research Program Manager, Microsoft Research