Faculty Summit 2012

Speaker: Tony Hey, vice president, Microsoft Research Connections

Speaker: Eric Horvitz, distinguished scientist and deputy managing director, Microsoft Research

Chair: Scott Counts, Microsoft Research
Speakers:

• Sharad Goel, Microsoft Research | slides
• James Pennebaker, University of Texas at Austin | slides
• Zeynep Tufekci, University of North Carolina at Chapel Hill | slides

This session addresses the study and use of social media. Social media have massively democratized the production and distribution of information, and as such are impacting virtually every aspect of our world, from politics and the economy to news and even crisis management. For scientists who want to study this phenomenon, social media represent a body of information that is real-time, largely text-based, semi-structured, highly diverse, and embedded in social structures. Thus, social media are of interest to a tremendous range of scientists, including ethnographers, sociologists, computer scientists, and computational linguists. This session will focus on techniques, application areas, and research opportunities in the social media domain.

Chair: Yinglian Xie, Microsoft Research
Speakers:

• Michael Reiter, University of North Carolina at Chapel Hill | slides
• Robert Sim, Microsoft | slides
• XiaoFeng Wang, Indiana University Bloomington | slides

Due to the increasing trend of moving user data and applications to the cloud, Internet service security is more critical than ever to the end user and large-scale services. This session examines various threats that target online services, ranging from user creation and hijacking attacks to those that abuse web and online advertising infrastructure, as well as sophisticated side-channel attacks in cloud-computing environments. We also discuss new directions and opportunities for attack detection and defense.

Chair: A.J. Brush, Microsoft Research
Speakers:

• Nilanjan Banerjee, University of Arkansas | slides
• George Demiris, University of Washington | slides
• Ratul Mahajan, Microsoft Research | slides

Homes contain an ever-increasing number of new devices and sensors, such as set-top boxes, game consoles, wireless routers, home automation devices, tablets, smartphones, and security cameras. The heterogeneity and complexity of technical innovations frustrate even technically savvy users and make it difficult for the home-computing infrastructure to meet its full potential. Join researchers from Microsoft Research and the academic community as they discuss technical challenges and research approaches for novel sensing in homes, as well as how to simplify the management and deployment of interesting new applications and devices in homes.

Chairs: Rane Johnson, Microsoft Research; Lucy Sanders, National Center for Women & Information Technology
Speakers:

• Magdalena Balazinska, University of Washington
• Emma Brunskill, Carnegie Mellon University
• Miriah Meyer, University of Utah
• Wei Wang, University of North Carolina at Chapel Hill

Microsoft Research has recognized innovative, promising new faculty members from a number of research institutions as Microsoft Research Faculty Fellows. More than 40 academic researchers whose exceptional talent for research and innovation identifies them as emerging leaders in their fields have joined the ranks of Faculty Fellows. The selected professors are exploring cutting-edge, high-impact research that has the potential to help solve some of the most challenging problems facing society today. In this session, three of the women Faculty Fellows discuss what they are doing to change computer science research and inspire the next generation of computer scientists. Hear how their universities helped them become successful and what more is needed to grow the next generation of women to become emerging leaders in computer science. Learn about possible keys to success and what all professors, researchers, and deans in computer science can do to help women become successful computer science faculty and researchers.

Chair: Stewart Tansley, Microsoft Research
Speakers:

• Patrick Baudisch, Hasso Plattner Institute | slides
• Gabe Cohn, University of Washington | slides
• Gerardo Gonzalez, Lancaster University | slides
• Joseph LaViola, University of Central Florida | slides

This session presents a view into Microsoft Research’s natural user interface (NUI) research portfolio by showcasing four partnerships, two in Europe and two in the United States. The presentations will highlight collaborative work on the following projects:

• Recent work toward scaling multitouch at whole room scales
• Research on using the human body as an antenna, exemplifying how touchless gesture interfaces could become ubiquitous
• A specific application of touchless gesture interfaces in surgical settings
• A pen-based prototype diagramming tool that uses constraint inference and a novel beautification algorithm to enable the drawing of precise geometric diagrams

While just a sampling of research in NUI today, this session aims to illustrate the diversity of the field and to inspire attendees by the fascinating opportunities and challenges in this field of research.

Chair: Patrice Godefroid, Microsoft Research
Speakers:

• Ben Livshits, Microsoft Research | slides
• Dawn Song, University of California, Berkeley | slides

Malware detection and analysis remains a rich topic of research as well as a significant challenge to consumers worldwide. This session focuses on detection and analysis as well as counter-measures to protect against malware.

Chair: Ece Kamar, Microsoft Research
Speakers:

• Luis Von Ahn, Carnegie Mellon University | slides
• Yiling Chen, Harvard University | slides
• Eric Horvitz, Microsoft Research
• Rajesh Patel, Microsoft | slides

In recent years, human computation and crowdsourcing have emerged as an important research area in computer science. Crowdsourcing uses human intelligence to solve tasks that computers cannot easily do alone. The work in this area encompasses investigations on different applications of crowdsourcing, best practices for crowdsourcing, and formal approaches for building more efficient and reliable crowdsourcing systems. Crowdsourcing research, which could result in easy and programmatic access to human intelligence, is of interest to a wide audience from different areas of computer science. In this session, leading researchers in crowdsourcing exchange information about recent developments and best practices, and discuss opportunities for future research.

Chair: Scarlet Schwiderski-Grosche, Microsoft Research
Speakers:

• Bill Gaver, University of London – Goldsmiths
• Albrecht Schmidt, University of Stuttgart | slides
• James Scott, Microsoft Research | slides

“People who are really serious about software should make their own hardware” – Alan Kay

This session will look at the importance of custom devices for research in domains such as ubiquitous computing, human-computer interaction, and the Internet of things. Recent advances have made it increasingly possible for researchers to build complex devices easily and rapidly. By using platforms such as Microsoft .NET Gadgeteer and taking advantage of physical prototyping machines (such as 3-D printing), one can now make fully custom devices (hardware, software, and physical form factor) rapidly and without expert domain knowledge. This enables researchers to deploy and evaluate systems more easily “in the wild.”

Chair: Hrvoje Benko, Microsoft Research
Speakers:

• Mark Bolas, University of Southern California | slides
• Steven Feiner, Columbia University | slides
• Jaron Lanier, Microsoft Research

Overlaying computer-generated graphics on top of the real world to create a seamless spatially registered environment is a core vision behind augmented reality (AR) research. While AR research exploded in the ‘90s with the promise of rich interfaces with head-worn displays, much of it never materialized beyond military uses. However, in the last five years, we have finally started to see AR technologies in consumers’ hands, albeit in a completely different form: smartphones equipped with cameras. Furthermore, Microsoft Kinect has opened up a whole range of AR experiences by simplifying camera-based tracking and gesturing. This session will share perspective on the current state of AR and offer insights of where AR might be going in the future.

Chair: Josh Benaloh, Microsoft Research
Speakers:

• Josh Benaloh, Microsoft Research | slides
• Lillie Coney, Electronic Privacy Information Center | slides
• Alex Halderman, University of Michigan | slides

Election technologies in current use have been shown to have significant vulnerabilities and are widely distrusted. In this session, experts on systems, data security, statistics, usability, and public policy will detail problems with current systems and examine approaches that can enable public verification of the accuracy of election results.

Chairs: Evelyne Viegas, Microsoft Research; Kuansan Wang, Microsoft Research
Speakers:

• Marti Hearst, University of California, Berkeley | slides
• Larry Heck, Microsoft | slides
• David Karger, Massachusetts Institute of Technology | slides
• Deborah McGuinness, Rensselaer Polytechnic Institute | slides

Next-generation information experiences dictate that the future embrace structured content and semantic inferences with increasingly interoperable data, services, and end-user personalization. This panel session of experts and thought leaders will explore the latest efforts in this direction and the research challenges to turn the vision into reality.

Chair: Nikolai Tillman, Microsoft Research | slides
Speakers:

• Christoph Csallner, University of Texas at Arlington | slides
• Suresh Lodha, University of California, Santa Cruz

In 2012, we are in the middle of a technology shift: instead of using PCs and laptops, mobile devices are becoming more prevalent for most everyday computing tasks. In fact, never before in human history were incredibly powerful and versatile computing devices such as smartphones available and adopted so broadly. In this session, we will discuss whether and how end-users can program mobile computing devices for fun and to automate tasks and implications for teaching computer science in today’s classrooms. Microsoft Research created the novel application creation environment TouchDevelop, which enables end users to author programs on smartphones—without a PC. This session will compare the traditional approach of using established programming languages to write mobile applications on a PC to novel programming approaches that enable creating applications on mobile devices directly. By mining existing mobile programming environments, we can infer what kind of applications end-users want to create and what challenges they face.

Chair: Paul Mitchell, senior director, Microsoft
Speakers:

• Mary Saunders, director, Standards Coordination Office, U.S. Department of Commerce, National Institute of Standards and Technology
• Sally Shipman Wentworth, senior manager of public policy, The Internet Society

How the Internet should be governed has become a hot topic in recent months. Some say that the current model of governance is unsustainable and should be changed. Others vigorously defend the current model as the enabling ingredient for unbounded innovation. Today, governments, academics, civil society organizations, and technologists are all engaged in the discussion. One thing they all agree on: “It is important to get it right!” But what is right? This session will examine the landscape and prospects for the next chapter in Internet history.

Speaker: Rick Rashid, chief research officer, Microsoft Research

Breaking down the barrier between the real world and the virtual world is a staple of science fiction, but technology is coming closer to actually blurring the line. Rick Rashid, Microsoft chief research officer and head of Microsoft Research, will present new, exciting hardware and software advances that are contributing to a simpler, more natural integration of the physical and virtual worlds.

Speaker: Harold Javid, director, regional programs, Microsoft Research

Computing increasingly enables the introduction of new products and new ways to work together, resulting in a growing requirement for skills, knowledge, and research capabilities that are the domain of post-doctoral computer scientists. Universities are the primary catalysts for encouraging and developing a new generation of computer scientists; companies also have a role to play. This short presentation examines some Microsoft programs and discusses features that can help graduate students and early-career academics advance their research capabilities and careers.

Chair: Donald Brinkman, Microsoft Research
Speakers:

• Anne Balsamo, University of Southern California | slides
• Andries van Dam, Brown University | slides
• Ethan Watrall, Michigan State University | slides

Increasingly, being able to visualize large collections of data is absolutely vital in the domain of cultural heritage—both for scholarly work and public consumption. Recent work explores novel and engaging ways to visualize and explore cultural heritage data collections. See demos of applications built on Microsoft PixelSense and tablets that demonstrate the advantages of pen and touch computing for enhancing the user experience of specific applications. These range from handwriting mathematics with pen or finger as input, to mathematical solvers, to using multitouch interaction for exploring large-format artworks in their rich context of related artworks, annotations, and guided tours. We will also explore the technological and social challenges of creating interactive exhibits around the AIDS Memorial Quilt, the largest community-created piece of folk art in the world. These applications are a beachhead on the untouched shores of big humanities research. They are what we can actually expect today and in the near future and are also first steps towards what might define an ideal user interface.

Chairs: Christian Bird, Microsoft Research; Thomas Zimmermann, Microsoft Research
Speakers:

• Ahmed Hassan, Queen’s University | slides
• Robin Moeur, Microsoft Research, and Ranga Narasimhan, Microsoft | slides
• Cameron Turner, Microsoft

Software development generates an inordinate amount of data. For example, check-ins, work items, and test executions are recorded in software repositories such as CVS, Subversion, GIT, and Bugzilla. Telemetry data reflects how customers experience software, which includes application and feature usage and exposes reliability. The sheer amount is impressive: for the 10,000 projects monitored by the web-page CIA.vc, every 17 seconds a commit takes place. The open-source platform SourceForge.net hosts more than 300,000 projects, and according to Github.com, 1 million people host 2.9 million GIT repositories. The bug database of the Mozilla Firefox projects now contains almost 700,000 reports, according to Ohloh.net. Microsoft has many sources of development data at similar scale. But how can this data be used to improve software projects? Software development analytics takes this data and turns it into insight to inform better development decisions. Analytics is commonly used in many businesses—notably in marketing, to better reach and understand customers. The application of analytics to software development is becoming more popular. In this session, we present successful efforts from Microsoft and academia that turn software development data into actionable insight.

Chair: danah boyd, Microsoft Research
Speakers:

• danah boyd, Microsoft Research
• Eduard Hovy, University of Southern California
• Mark Latonero, University of Southern California
• Mary G. Leary, Catholic University of America
• Janis Wolak, University of New Hampshire

Networked technologies—including the Internet, mobile phones, and social media—alter how information flows and how people communicate. There is little doubt that technology is increasingly playing a role in the practices and processes surrounding human trafficking: the illegal trade of people for commercial sexual exploitation, forced labor, and other forms of modern-day slavery. Human trafficking has many facets to it and technology’s role varies as a result. Yet, little is known about costs and benefits of technology’s role. We do not know if there are more human trafficking victims as a result of technology, nor do we know if law enforcement can identify perpetrators better as a result of the traces that they leave. We do know that technology makes many aspects of human trafficking more visible and more traceable, for better and for worse. Focusing on whether technology is good or bad misses the point; it is here to stay, and it is imperative that we understand its role. More importantly, we need to develop innovative ways of using technology to address the horrors of human trafficking. This panel will explore ways in which technologists can use data traces to help address challenges in the human trafficking arena, paying particular attention to the commercial sexual exploitation of children. We will discuss ways in which technology can help identify perpetrators or victims, empower law enforcement, and disrupt organized crime. We will also examine the complex tensions involved in balancing the need to combat child victimization with the importance of civil liberties and privacy. Finally, we will discuss the possibilities of—and need for—technical innovation that can counter human trafficking abuses.

Chair: Stewart Tansley, Microsoft Research
Speakers: Surya Vanka, Microsoft

The tile-based user interface (UI) first showcased for Windows Phone has become the basis of the new look and feel of the experience of most of Microsoft’s products. But this is more than a UI language. It is a set of design principles that have been adopted across the company. What’s the background to this approach? In this presentation, Vanka will talk about the user experience revolution at Microsoft and the emergence of modern experiences. He will describe why this approach represents a significant shift away from the stylistic UI approaches that are prevalent across the software industry, and will discuss implications of increased attention that this approach is receiving from the academic research community.

Chair: Juan Vargas, Microsoft
Speakers:

• Krste Asanović, University of California, Berkeley | slides
• Tim Mattson, Intel | slides
• Josep Torrellas, University of Illinois at Urbana-Champaign | slides

For decades, the microprocessor industry improved the performance of on-chip processors by increasing the number of operations executed per CPU cycle. By 2007, it was clear that heat and power would limit further increases in clock rates. Today, it is widely accepted that performance gains can only be obtained by putting more than one processor in a single chip. Multicore parallelism represents a paradigm shift that breaks the 50-year tradition of conventional sequential programming models and software development. Knowing that research in this area would have deep implications, Microsoft and Intel launched in 2008 the Universal Parallel Computing Research Centers (UPCRC) at the University of Illinois at Urbana-Champaign (UIUC) and UC Berkeley. Through this unique collaboration, Microsoft, Intel, UIUC, and UC Berkeley have joined forces to conduct research into all aspects of parallelism at the client, including architectures, operating systems, run-time middleware, programming tools and libraries, and applications. During this session, researchers from Microsoft, Intel, UC Berkeley, and UIUC will share the outcomes of their three-year efforts and will present their insights about new directions in the changing landscape of parallel computing.

Microsoft Research has numerous programs for supporting students and academics. During this breakout session, the people who run the programs—from internships to faculty fellowships, innovation funds to woman’s scholarships—will be on hand to discuss how the programs work and how you can participate.

Chair: Jaime Teevan, Microsoft Research
Speakers:

• Lili Cheng, Microsoft Research | slides
• Cliff Lampe, University of Michigan | slides
• Meredith Ringel Morris, Microsoft Research | slides

Social search is an emerging research area that explores how social interactions and social data can enhance existing information-seeking experiences, as well as enable new information retrieval scenarios. This session will showcase different models of social search, including 1) the use of social data to augment search, 2) social data as new information to be searched, and 3) social interaction and collaboration as part of the search process.

Chair: Kenji Takeda, Microsoft Research
Speakers:

• Manuel Chakravarty, University of New South Wales | slides
• Antonio Cisternino, University of Pisa | slides
• Don Syme, Microsoft Research | slides

This session highlights some of the latest developments and applications of functional programming. Information-rich programming is changing how we think about language design, with F# type providers bringing information to developers over the web. Discover how Haskell is being extended into parallel and cloud computing. And learn how functional programming is being used in the wild, from bioengineering and chemistry to engineering and finance. This session will show how functional programming can transform business, science, and engineering in the real world, for today and tomorrow.

Chair: Rane Johnson, Microsoft Research
Speakers:

• Walter Alvarez, University of California, Berkeley
• Sergey Berezin, Moscow State University | slides

Imagine a world where scientists, researchers, students, and teachers collaborate to share historical information through images, videos, documents, charts, interactive tours, and more. Imagine a world where leading academics publish their findings to the world in a manner that can easily be accessed and compared. ChronoZoom is an open-source community project dedicated to visualizing the history of everything and supporting the emerging field of Big History. Big History is the attempt to understand, in a unified, interdisciplinary way, the history of the cosmos, Earth, life, and humanity. ChronoZoom seeks to bridge the gap between the humanities and sciences by enabling all kinds of information to be visually presented and organized. In this session, learn about the challenges of teaching a course on Big History, and what it is like to work on a multi-disciplinary project with humanistic, scientific, and computer science researchers. Understand what ChronoZoom hopes to accomplish with the research community and how computer scientists can better work with scientists and humanists. Get a behind-the-scenes look at how we brought ChronoZoom to life through HTML5 and Windows Azure. We will address the various visualization challenges, data management issues, and user interface questions solved in this project and the complex algorithms created. We will also discuss what we hope to accomplish in the next phases of ChronoZoom development and how other computer science researchers can work with the team.

Chair: Juan Vargas, Microsoft | slides
Speakers:

• James Demmel, University of California, Berkeley | slides
• David Padua, University of Illinois at Urbana-Champaign | slides
• David Patterson, University of California, Berkeley
• Burton Smith, Microsoft | slides

In this session, Microsoft Technical Fellow Burton Smith will present “Operating System Architecture for Parallel and Distributed Computing.” In addition, a panel will discuss the educational impact of the Universal Parallel Computing Research Centers (UPCRC) and deliver the following presentations:

• What to Cover in an Introductory Course on Parallel Programming?—David Padua
• Teaching Parallel Computing Nationwide and Beyond via NSF/XSEDE—Jim Demmel
• Teaching Software Engineering to 35,000 Students Using a Massive Open Online Course (MOOC)—David Patterson

Speaker: David Breashears, founder and executive director, GlacierWorks

The Himalaya Mountains are home to the world’s most magnificent peaks and thousands of high-altitude glaciers. These important glaciers supply crucial seasonal flows to rivers across Asia, yet many are disappearing at an increasing rate. In this session, mountaineer and photographer David Breashears presents his recent photographs of the world’s least studied glaciers. By comparing them to archival photographs taken over the past century by the world’s greatest alpine photographers, the alarming loss of ice is starkly revealed.

Since 2007, GlacierWorks has embarked upon 11 expeditions to the greater Himalayas. Breashears will describe the increasingly sophisticated imaging technology that the GlacierWorks team has employed to capture high-resolution photography of a wide range of glaciers and their surrounding geography—resulting in terabytes of high-resolution images from hundreds of strategic locations throughout the greater Himalayan region. These images, along with terabytes of images captured during a series of helicopter flights around Mount Everest, are being built into the highest resolution 3-D point cloud ever created of the Himalayan region.

GlacierWorks’ extensive imagery, along with archival media, is currently being incorporated into an immersive Internet experience, which uses Rich Interactive Narratives technology from Microsoft Research. GlacierWorks hopes, through interactive and engaging educational experiences such as this, to interest students in gaining a better understanding of climate change’s impact on the glaciers and river systems of Asia.