External Research Overview Tony Hey, Corporate Vice President, External Research, Microsoft Research

Presentation: External Research Overview

Rick Rashid, Senior Vice President, Microsoft Research (Moderator) Daniel A. Reed, Director of Scalable and Multicore Computing, Microsoft Research Edward W. Felten, Professor of Computer Science, Princeton University Howard A. Schmidt, President and CEO of R & H Security Consulting LLC Elizabeth Lawley, Director of the Laboratory for Social Computing at the Rochester Institute of Technology

Information gathering, personal and business communication, social interaction, entertainment and learning increasingly flow into a cyberspace interconnected universe. Along with the broad benefits and improvements in our knowledge of the world we live in come new challenges for individuals and society, ranging from security and privacy to effective uses of new modes of interaction. Information overload has achieved new levels. As a result, this session will examine the current state of this interconnected universe, its implications, future needs, and corresponding research opportunities.

Presentation: The Cyberspace Connection – Impact on Individuals, Society, and Research

Video podcasts, live-blogging, data visualization, and the wisdom of crowds: who needs publishers and peer review anymore? Well, we all do, and we probably will for years to come. So, while the sociology of academia shifts and as new business models evolve, it is worthwhile to pay close attention to the variety of trends and breakthroughs across various disciplines and the impact they will have on scientific dissemination overall. The role of the researcher, the publisher, and the repository/library are shifting dramatically to accommodate data and other digital content, as well as the concept of interoperability. These talks will highlight some innovative projects and the thought-leaders who are helping to re-architect the scholarly world as we have come to know it.

Presentation: Lee Dirks, Introduction

Presentation: Paul Ginsparg, Scholarly Communications

Presentation: Natasa Milic-Frayling, Vijay Rajagopalan, PLANETS & Document Interoperability

Technical and infrastructural developments have come together to create a fertile environment for developing and fielding AI applications that promise to provide value to people in the course of their daily lives. These developments include (1) technical advancements in machine learning and reasoning, (2) the growth in CPU and memory capabilities within commonly available devices and platforms, (3) the connectivity, content, and services provided by the evolving Web, and (4) the increasing availability of data resources, including corpora of behavioral data collected via inexpensive sensors, and through ongoing interaction with software and services. I will discuss several trends in AI research and reflect about problems and opportunities ahead. I will highlight ideas in the context of active research projects at Microsoft Research and other centers.

The goal of Microsoft’s Parallel Computing Platform (PCP) team is to enable the shift to modern, multi- and manycore hardware, by providing a runtime, programming models, libraries, and tools that make it easy for developers to construct correct, efficient, maintainable, and scalable programs through the use of parallelism. In doing so, tens of years of industry research has been combined and applied in a myriad of ways. This talk examines PCP’s current progress, explicitly relating it to specific research of the past and present, in addition to surveying future efforts and possible research opportunities.

Presentation: Joe Duffy, Microsoft’s Parallel Computing Platform: Applied Research in a Product Setting

The talk will report on various research and teaching projects carried out by the Operating Systems and Middleware group at HPI that are based on Phoenix and the Windows Research Kernel (WRK). Phoenix is an extensible compiler (construction kit) developed by Microsoft. We have used Phoenix to implement Gripper-Loom.NET, our implementation of Aspect-Oriented Programming for the .NET framework. AOP and Phoenix also form the foundation of our Dynamic Software Update Platform (DSUP). Applicability of DSUP has been demonstrated with various freely available .NET programs (Paint.NET, LumiSoft imap/ftp) as well with the Distributed Control Lab, our tele-laboratory environment. The Windows Research Kernel contains (most of) the kernel sources of Windows Server 2003 R2. We have extended the WRK build environment for classroom use (Visual Studio, Phoenix). Besides just using Phoenix to build the kernel (Apr 2008 CTP and Jul 2007 RDK), we have extended Phoenix to allow for generation of hyperlinked sources and documentation of the WRK. The talk will report about experiences with WRK-based labs (like implementing system calls, changing scheduler) in context of the Operating Systems curriculum at HPI. With the Kernel-level Scheduling Server, the Windows Monitoring Kernel, and the kstruct-approach towards system monitoring, the talk will also give an overview over current research projects on operating system support for service computing.

Presentation: Andreas Polze, Phoenix and the Windows Research Kernel: Enhancing the Teaching/Learning Experience

This presentation will focus on how the subtle style of Chinese water color painting can enhance game design, demonstrating how beauty and serene play enhanced by peaceful music can provide a unique and delightful game experience. The Animation Institute of the Beijing Film Academy is exploring a new application pattern of Classic Chinese Culture in modern entertainment and life. This project shows the beauty of Classic Chinese Painting in a digital way, by exploring how to make the planed Chinese Ink Painting into a three-dimensional dynamic display. It also seeks to make ancient culture more entertaining and easily accepted by the younger generations.

Presentation: Sun Li, Applying Chinese Water Color Painting Style to Game Design

We are in the midst of a generational shift in research, driven by new “disruptive” technologies. The rapid emergence of a new world of eScience driven by very large scale data, next generation sensors, and advanced robotic instruments, in a host of disciplines from environmental, physical and other sciences and engineering through public health and medicine, requires a new approach to provide computational and data management tools and services for research. This session will focus on a software + service model with scientific services delivered from the Cloud that we believe will become an increasingly accepted model for research.

Presentation: Roger Barga, Introduction

Presentation: Ian Foster, Towards a Data Cauldron

Presentation: Dennis Gannon, The Scientific Data Center

Presentation: Paul Watson, The CARMEN Science Cloud & Beyond

Artificial Intelligence

A. AI Meets Markets: Trading Agents and Strategic Reasoning Michael P. Wellman, University of Michigan

Trading agents are autonomous computer programs that bid in electronic markets without direct human intervention. The design and analysis of trading agents is a challenging problem for artificial intelligence, but one with great potential for electronic commerce. Through a series of international trading agent competitions, AI researchers have exposed lessons and principles for effective trading agent design. Consideration of market environments has pushed the limits of computational game-theoretic technique, leading to advances in strategic reasoning based on simulation, machine learning, and other empirical methods.

Presentation: Michael P. Wellman, AI Meets Markets: Trading Agents and Strategic Reasoning

B. AI, Sensing, and Optimized Information Gathering: Trends and Directions Carlos Guestrin, Carnegie Mellon University

Over the last few years, we have tackled a fundamental problem that arises when using sensors to monitor the ecological condition of rivers and lakes, the network of pipes that bring water to our taps, or the activities of an elderly individual when sitting on a chair: Where should we place sensors in order to make effective and robust predictions? In this talk, I will outline some of the algorithmic and theoretical developments that enabled these applications; drawing an interesting connection between sensor placement for water monitoring, and the problem of selecting blogs to read in order to learn about the biggest stories discussed on the web. I will then outline a more general, longer-term agenda for using information gathering to drive new, broader research in AI. This talk is primarily based on joint work with Andreas Krause.

Presentation: Carlos Guestrin, AI, Sensing, and Optimized Information Gathering: Trends and Directions

In March 2008, Microsoft and Intel Corporation announced their partnering with academia to create two Universal Parallel Computing Research Centers (UPCRC), aimed at accelerating developments in mainstream parallel computing in desktop and mobile computing. The research centers are located at the University of California, Berkeley (UC Berkeley) and the University of Illinois at Urbana-Champaign (UIUC). Research will focus on advancing parallel programming applications, architecture and operating systems software. This is the first joint industry and university research alliance of this magnitude in the United States focused on mainstream parallel computing. This session will present an overview of the UPCRC partnership. Dan Reed will discuss challenges and opportunities presented by the shift to multicore/manycore computing, and David Patterson and Marc Snir, the directors of the two Centers, will present their UPCRC research agendas.

Presentation: Dan Reed, Universal Parallel Computing Research Centers (UPCRC)

Presentation: David Patterson, The Parallel Computing Landscape: A View from Berkeley

Presentation: Marc Snir, Making Parallel Programming Synonymous with Programming

There are currently several different approaches to semantics, semantic technologies, and the Semantic Web floating around. While the uptake of these technologies is going well, there is still confusion about what sort of technology fits where and how it works. The confusion is made worse because the term “ontology” is used in a number of different ways. In this talk, James Hendler will describe how different sorts of models can be used to link data in different ways. He will particularly explore different kinds of Web applications, from Enterprise Data Integration to Web 3.0 startups, and the different kinds of techniques needed for these different approaches.

Introduction: Evelyne Viegas, Intelligent Web

Presentation: James Hendler, Ontologies on the Web

The younger generation is showing us the way in how we stay in touch, interact, and exchange information. Social networking sites are all about creating machine processable structures, networks of connections from which information can be inferred. The “wisdom of the crowds” is being applied to identify trends, and it is because of advances in technology that this is now possible. It’s no surprise that the research community is paying attention. Could we leverage the same ideas in order to advance the way we collaborate, exchange research information, review the works of our peers? And, if we can represent social connections using machine representations, could we do the same for all possible data-to-data connections? What is the role of semantic computing? What is the role of the Semantic Web? How do we bring everything together?

Presentation: Noshir Contractor, Enabling Social Networks Using Semantics

Presentation: David De Roure, myExperiment

Presentation: Jennifer Golbeck, Social Networks on the Semantic Web

Presentation: Marc Smith, Social Networking and Semantics

A. Toward Situated Interaction Dan Bohus, Microsoft Research

The Situated Interaction project at Microsoft Research aims to enable a new generation of interactive systems that can reason about their surroundings and embed interaction deeply into the natural flow of everyday tasks, activities and collaborations. As an initial sample challenge in this space, we are currently developing a situated conversational agent that can act as a Microsoft front-desk receptionist (e.g., making shuttle reservations, registering visitors, providing campus information). The system integrates a large number of AI technologies such as speech recognition and language understanding, face detection and tracking, intention recognition, engagement and behavioral modeling in a conversational framework that allows it to engage in mixed-initiative natural language interaction with one or multiple participants. In this talk, I will briefly outline the Situated Interaction project, describe and demonstrate the current Receptionist system and discuss some of the research challenges under investigation: multi-participant engagement and dialog models, conversational scene analysis, spatio-temporal trajectory reasoning, and behavioral modeling.

Presentation: Dan Bohus, Toward Situated Interaction

B. Statistical Machine Translation Research at Microsoft Research Robert Moore, Microsoft Research

Over the last 20 years, statistical methods have revolutionized the 60-year quest for automatic translation between natural languages. This talk will review some of Microsoft Research’s recent accomplishments in this burgeoning field, placing them in the context of the overall state of the art.

Presentation: Robert Moore, Statistical Machine Translation Research at Microsoft Research

C. Interactive Machine Learning: Challenges, Methods, and Applications Ashish Kapoor, Microsoft Research

I will present research on interactive machine learning, principles and applications of methods that involve human in the loop and adapt continuously, so as to learn effectively over prolonged periods of time. I will define a set of challenges within this realm, including the problem of triaging computational abilities and human effort by identifying the most important action to perform under scarce informational resources, handling the potential non-stationarity of environments, and designing learning systems to work effectively with people-as many applications can benefit by considering humans in the loop. I will discuss principles in the context of several applications, including voicemail management, image categorization and search, and experience sampling for building predictive user models.

Presentation: Ashish Kapoor, Interactive Machine Learning: Challenges, Methods, and Applications

The main intellectual difficulty of concurrent programming lies in reasoning about the interaction between concurrently executing threads. Nondeterministic thread scheduling makes it extremely difficult to reproduce behavior from one run of a program to another. As a result, the process of testing and debugging concurrent software becomes tedious, resulting in a drastic decrease in the productivity of programmers and testers. These talks will present new research in the area of analysis and testing tools for concurrent programs.

Presentation: Madan Musuvathi, Shaz Qadeer, CHESS: Systematic Concurrency Testing

Presentation: Koushik Sen, Active Random Testing of Parallel Programs

Presentation: Yuanyuan Zhou, Learning from Mistakes – Real World Concurrency Bug Characteristics

Web search presents incredible opportunities and challenges for text analysis. Billions of documents, queries, and clicks, are collected each day. Tens of thousands of computers are available, to extract the signal from the noise. At the same time, documents and queries must be processed in milliseconds with high precision. Paul Viola will present some of the Microsoft Live Search work on the analysis of queries and documents, which attempts to navigate these waters to extract value from the deluge.

Introduction: Evelyne Viegas, Intelligent Web

Presentation: Paul Viola, Information Extraction from Documents and Queries

One of the most powerful tools for improving success rates in introductory computing courses is the incorporation of context—a theme that pervades the computing lectures, assignments, and examples, which relates the content to a concrete application domain. Contextualized computing education has even allowed us to be successful with challenging audiences, such as the non-technical major. In this talk, we review why the Georgia Institute of Technology has chosen to teach serious computer science to every student on campus, and then discuss research findings from multiple schools on the benefits and costs of contextualized computing education.

Presentation: Mark Guzdial, Contexts in Computer Science Education

Scientific discovery business practice and social interactions are moving rapidly from a world of homogeneous and local systems to a world of distributed software, virtual organizations, and cloud computing infrastructure. In computing, a tsunami of new experimental and computational data and a suite of increasingly ubiquitous sensors pose new opportunities for data analysis, transport, visualization, and collaboration. In society and business, software as a service and cloud computing are empowering distributed groups in new ways. This talk will describe a vision of Microsoft’s clouds that can be a platform for academic computing research and possible partnerships.

Presentation: Dan Reed, The Future of Research Clouds

Demonstrations: Dan Bohus, Microsoft Research; Ken Hinckley, Microsoft Research; Zhengyou Zhang, Microsoft Research

Presentation: The Latest Technologies From Microsoft Research

Presentation: Dan Bohus, Ken Hinckley, Situated Interaction

Presentation: Zhengyou Zhang, Personal Telepresence Station

Network science is a growing interest in many disciplines. New tools from Microsoft Research make performing many basic network manipulations and visualization tasks as simple as using Excel. The (Excel) .NetMap add-in provides directed graph charting features within Excel, allowing users to create node-link diagrams with control over each node and edge color, size, transparency and shape. Since .NetMap builds within Excel, all of the controls and programmatic features of Office are available. Additional features of (Excel) .NetMap generate social networks from data sources like personal e-mail (drawing data from the Windows Desktop Search engine). Arbitrary edge lists (anything that can be pasted into Excel) can be visualized and analyzed in .NetMap. This session will provide a walk through the basic operation of .NetMap. Attendees are encouraged to bring an edge list of interest. Sample data sets will be provided. To download the Excel .NetMap Add-in and slides, go to the following Web site: here.

Presentation: Marc Smith, (Excel) .NetMap A Toolkit for Social Network Analysis

While most faculty want to understand the impact of their teaching on students, few have the time or resources to engage in in-depth study to become assessment experts. REAssess: Resources for Educational Assessment is a new interactive assessment Web site that focuses on the computing, science, technology, engineering and math disciplines. The simple tools and resources included will help take the drudgery out of conducting educational assessment and offer ways to gain new insights into student learning. This session will introduce the REAssess Web site and tools and offer some lessons learned and future directions for the project.

Presentation: Stephen T. Kerr, Steven Tanimoto, REAssess: Resources for Educational Assessment

The predictions around the advent of eScience are slowly becoming reality for the scientific community. However, the implications both near-and longer-term for society in general are still unclear. It seems obvious that the broad availability of cyber-infrastructure will have a impact on different non-scientific populations, but exactly what and when? How does one become a Citizen-Scientist? How will we use/access eScience resources without even knowing it? This panel of experts will field this wide-ranging topic and provide their educated viewpoints on how eScience will soon blend into the very fabric of our everyday lives.

Presentation: Daron G. Green, The Broader Impact of eScience

Presentation: Peter Lee, eScience: Data, Computing, and Crowds

Presentation: Adam Siepel, Implications of eScience for Science and Society: A View from Genomics

Presentation: Wei Wang, eScience: Promoting Public Engagement

Presentation: Jeannette M. Wing, eScience for All: Not If, But When

This panel of leading researchers will discuss key challenges and opportunities in AI theory and practice, building on and extending topics covered in earlier breakout presentations, and bringing their own personal perspectives.

Webcast: Artificial Intelligence Theory and Practice: Hard Challenges and Opportunities Ahead

This panel will present the goals and important activities of four major organizations working, often in cooperation, to improve the representation of women in technology CRA’s Committee on the Status of Women in Computing Research (CRA-W), the National Center for Women and Information Technology (NCWIT), ACM’s Committee on Women in Computing (ACM-W) and the Anita Borg Institute (ABI). Panelists will discuss their programs and present opportunities for active involvement. Highlighted will be a number of projects in which Microsoft has partnered with the organizations.

Presentation: National Engagements for Promoting Women in Computing

This session will provide an overview of Microsoft Research Asia (MSRA), its organization, research areas, and accomplishments since its establishment in November 1998. As MSRA is approaching its 10th anniversary, Dr. Hon will provide a highlight of MSRA’s impact in academic community, product transfer, and his view for the future development. Lolan Song will give an overview of the academic collaboration programs in the Asia-Pacific region. There will be a brief Q&A session at the end of this talk.

Presentation: Hsiao-Wuen Hon, Microsoft Research Asia Overview

In his talk entitled “Intelligent Webs of Photos (By Accident and By Design),” Blaise Aguera y Arcas will begin with a brief history of the developments in computer vision leading up Photo Tourism and Photosynth, technologies that allow geometric relationships between photos to be inferred automatically and used as navigational affordances. These techniques can be applied to existing image corpi on the Web, with varying degrees of success. He will then consider what can be done with Photosynth as a collaborative and/or community-oriented authoring tool. He will cover some new techniques developed for enhancing the navigability of such data, and the different challenges posed by authored synths (or synths “by design”) and found or “accidental” synths.

Virtual Earth is a general platform for storing and accessing data with location information. In the second talk of the session, entitled “Geo-Positioned Media within the Context of Virtual Earth,” Eyal Ofek and Bill Chen will discuss the benefit of relating media within the context of Virtual Earth. They will address issues of geo-positioning media, its organization, and efficient storage and distribution by factoring repeating content.

Presentation: Bill Chen, Eyal Ofek, Geo-Positioned Media within the Context of Virtual Earth

Presentation: Evelyne Viegas, Bill Chen, Virtual Earth – Academic Research Collaboration 2007 RFP

This session will present a novel approach to the analysis and empirical evaluation of games that purport to be “games for learning,” with the goal of creating a predictive set of design principles that can be tuned to the desired user base, game genre, and area of learning.

Presentation: Ken Perlin, Games for Learning: Understanding What Makes an Effective Game for Learning

Presentation: Lili Cheng, Design Expo

This panel of distinguished researchers has focused on computer-mediated, collective action-systems where many people come together and interact through computers. Each has been researching and teaching network science in general and social network science in particular. This panel will focus on the concepts, methods, and Internet services that enable social networks have had an impact in their classrooms and research.

Presentation: Karrie Karahalios, Social Visualization: Communication, Signal, or Cue?

Presentation: Cliff Lampe, Using Social Network Sites in Education

This session will provide an overview of the research agenda in computer science in some Latin American countries, and a brief background about the organization and infrastructure of existing regional cooperation programs. Universities in the Latin America and Caribbean region are well positioned to become strategic partners in national innovation systems and strongly contribute to economic and social development. During this session, projects conducted by academic researchers in specific countries in Latin America will be discussed as examples of research success stories from this emerging region.

CmapTools: From Meaningful Learning to a Network of Knowledge Builders Alberto Cañas, Institute for Human and Machine Cognition (IHMC)

Based on the Theories of Meaningful Learning and Education of Ausubel and Novak, we have developed software tools that allow users to collaborate in the construction of shared knowledge models based on concept maps. These tools are used worldwide by users of all disciplines and ages, from elementary school students to NASA scientists. In this talk, we’ll discuss the software tools and how they are being used at the Proyecto Conectate al Conocimiento in Panama, an effort to transform public school education through appropriate use of technology that has led to a nationwide network of knowledge builders, and will describe the methodology on which this concept map-centered learning environment is based.

Pushing and Pulling Information From the Mexican Dataspace Genoveva Vargas-Solar, French Council on Scientific Research (CNRS), Informatics Laboratory of Grenoble, French Mexican Laboratory of Informatics and Applied Automatic Control

The increasingly global economy calls for continuous access to information in a flexible and robust way through services that come up as a new paradigm for programming and organizing operations. At the same time, the emergence of ubiquitous computing introduces wireless and portable technologies that democratize access to information and services and thereby opens new research challenges for querying techniques that can cope with this novel dynamic execution environment. Research on query processing is still promising given the explosion of huge amounts of data largely distributed and produced by different means (sensors, devices, networks, analysis processes), and the requirements to query them to have the right information, at the right place, at the right moment. This challenge implies composing services available in dynamic environments and integrating this notion into query processing techniques. Academic and industrial efforts must address novel challenges on data/services querying that go beyond existing results for efficiently exploiting data stemming from many different sources in dynamic and multi-scale environments. This talk will discuss the challenges of modern data and services intensive systems deployed on networks of heterogeneous devices, the so called ecosystem or dataspaces.

Presentation: Genoveva Vargas-Solar, Pushing and Pulling Information From the Mexican Dataspace

How Can Microsoft Help Geneticists? Mayana Zatz, Human Genome Research Center, University of Sao Paulo

Human geneticists face problems that require each time more robust computer support and science such as: building up genealogies with hundreds of members and linking them up using a friendly system; improving micro arrays analysis; organizing the gigantic amount of data that will be progressively generated by the human genome. Some examples will be given during the talk opening the way for questions and suggestions.

Presentation: Mayana Zatz, How Can Microsoft Help Geneticists?

As the 2007-2008 SensorMap RFP projects complete their work, we have invited a panel of researchers from some of these projects and others to discuss their experiences. These experiences include leveraging the SenseWeb/SensorMap platform to plan, deploy and monitor sensor experiments and analyze spatio-temporal sensor data, as well as lessons in integrating sensornet technologies with the Web and mapping tools such as Virtual Earth.

Presentation: Feng Zhao, Introduction

Presentation: Magdalena Balazinska, Event Detection and Notification in the World-Wide Sensor Web

Presentation: Tarek Abdelzaher, Privacy and the Participatory Sensor Web

Presentation: Sebastian Michel, Environmental Monitoring 2.0

Presentation: Liqian Luo, SenseWeb and SensorMap

Presentation: Prashant Doshi, Semantic Reconciliation of Sensor Net Meta-Data

Presentation: Evelyne Viegas/Bill Chen, Virtual Earth Academic Research Collaboration 2007 RFP

Over the past ten years, increased access to computing resources and expanded connectivity have contributed to greater opportunity for more people to participate in the information communication technology (ICT) ecosystem as users, creators, or both than ever before. During this time, both open source software (OSS) communities and Microsoft have made significant contributions and experienced substantial growth. Sam Ramji leads a cross-disciplinary, cross-company community establishing practices to foster constructive and complementary relationships between Microsoft and OSS. This talk will describe the principles underlying Microsoft’s open source business, technical, and community participation strategy, summarize key milestones to date, and share a vision for the relationship between Microsoft and open source communities in the future.

Presentation: Sam Ramji, Participation in a World of Choice: Open Source and Microsoft