Top students contemplate the shape of the AI future at PhD Summer School

Published

By , Director

Beautiful summer weather, ice cold Pimm’s and the idyllic scenery along the River Cam greeted over 100 PhD students from across the EMEA (Europe, Middle East and Africa) region who had traveled to Cambridge, England, to share and learn about artificial intelligence with tech visionaries and experienced researchers.

Each year the Microsoft Research Cambridge Lab brings together tech luminaries and researchers with PhD students from research institutions across EMEA to explore the state of the art in computing as well as get a better appreciation for all aspects of what it means to be a successful researcher, including the ability to storytell their research and to give consideration to the social, economic and even philosophical effects of advancing computing technology. This year’s PhD Summer School event was the second to focus on the topic of AI, aligning with Microsoft’s business vision.

Microsoft Research Podcast

AI Frontiers: Models and Systems with Ece Kamar

Ece Kamar explores short-term mitigation techniques to make these models viable components of the AI systems that give them purpose and shares the long-term research questions that will help maximize their value. 

Over 100 students from 32 research institutions spanning 11 countries in EMEA exemplified diversity going beyond nation and culture with research backgrounds extending beyond computer science and engineering into the realms of design and various natural and social sciences. Attendees included recipients of Microsoft Research PhD Scholarships, along with students working with EMEA Joint Research Centers, collaborating on Microsoft Azure for Research projects or otherwise partnering with Microsoft Research Cambridge.

Socio-technical Challenges

The breadth of topics covered at this year’s Summer School was impressive, spanning everything from the purely technical to practical application of AI, to honing researcher skills, and ethics. Students spent a week diving into deep learning, reinforcement learning, computer vision, and security sessions in addition to others devoted to causality, security and model-based machine learning.

Other sessions focused on ethics in AI and the human consequences of AI on societies. Of particular interest to the students were the keynotes and this year’s speakers didn’t disappoint, delivering thoughtful and challenging addresses that focused on the socio-technical challenges of AI. Jacqueline Krones, Principal Design Researcher at Microsoft spoke on “Using Ethical Principles to Inform Industry Solutions” that presented Microsoft’s perspectives and principles around AI development, looking, for example, at different attitudes globally around issues of trust and privacy. AI systems are clearly having a huge impact on our lives and they raise many questions both in terms of ethics (avoiding data bias, being transparent about how they make decisions) and the requirement they put on technologists to develop their research and product concepts in a way that is responsible and transparent. In an inspiring keynote, Dame Wendy Hall, Regis Professor of Computer Science at the University of Southampton and recently appointed Skills Champion for AI in the UK (advising the UK government on how the artificial intelligence industry can be grown in the country) spoke at length. Hall drew parallels between the development of the internet from its early years, and the burgeoning growth of AI systems that is taking place now, underscoring a message emphasized throughout the program – that a socio-technical perspective is crucial when innovating in the AI space, and that working with other disciplines is key to understanding the systems researchers are building in order to mitigate negative effects.

“AI Through the Looking Glass” by Dame Wendy Hall

Talks

On the more technical side, MS Technical Fellow and MSRC Lab Director Christopher Bishop talked about model-based machine learning, giving a historic perspective that drew from his almost two decades of machine learning work at Microsoft Research Cambridge. Hermann Hauser, a Cambridge-based serial entrepreneur and co-founder of Amadeus Capital Partners, spoke about “Intelligent Machines”, comparing the biological intelligent machines that are human beings with the artificial intelligence machines created by them and emphasizing the importance of choosing human-oriented goals as AIs increase in power and ubiquity.

Healthcare sessions were also a huge draw at the event, featuring engaging talks by MS Principal Researcher Jasmin Fisher who talked about prediction of drug resistance based on in silico cancer programs, Mihaela van der Schaar of the University of Oxford who introduced AutoPrognosis, a suite of tools that can help craft machine learning models for a variety of diseases, and Rajesh Jena of Addenbrooke’s Hospital Cambridge and a part-time researcher at Microsoft Research Cambridge who highlighted practical ways in which machine learning can improve outcomes for patients undergoing radio therapy. Finally, Kenji Takeda of Microsoft Research gave an extraordinary overview of the healthcare research carried out throughout Microsoft.

“The AI summer school was an incredible experience. It was lovely to be in Cambridge and Microsoft Research, I learned so much and it felt so good to be so well taken care of. I learned a lot of good things about MSR, as well as making some good connections at MSR and other companies nearby. Thank you very much, I’m telling everyone I know about it!” – AI Summer School Student

Steven Abrams demoing “Soundscape” and Haiyan Zhang demoing “Emma” at DemoFest

DemoFest

The Summer School really came to life at DemoFest where students learned about some of the most compelling developments coming out of the Microsoft Research Cambridge lab. Thirteen demos were presented to the students, covering research on everything from accessibility to Microsoft Hololens.

Poster Session

Another engaging highlight was the Poster Session where 21 PhD students presented posters of their research to their peers, MS researchers, and a number of visitors. The students and attendants enjoyed the face-to-face networking time, staying late into the evening well past the scheduled end time of the session. The success and duration of the poster session really highlighted the enthusiasm and passion displayed by all of the invited students at the week-long Summer School.

“I just had to thank you for organizing such a fantastic week! Not a morning nor afternoon went by when I didn’t learn something new and useful to put into practice. It was great hearing from, talking to and mingling with so many lovely and talented people. Thank you for putting it all together so beautifully,” said one AI Summer School Student.

Throughout the week, the theme returned to again and again by researchers and attendees was that of a self-awareness that these students of today would shape the future of tomorrow. “At Microsoft Research, we care deeply about the next generation of AI talent and that is why we take an approach that aims to preserve and nurture talent within academia, in a way that is both sustainable and mutually beneficial,” said Christopher Bishop.

A night off – pre-dinner drinks and games at Downing College, Cambridge

Related Items:

 

 

Continue reading

See all blog posts