Home | Agenda Day 1Agenda Day 2 | Bios DemoFestVideosImages | FAQ 

Microsoft Research Faculty Summit 2009

DemoFest Booths




WorldWide Telescope–Earth
See a demonstration of the amazing visualization features in Microsoft Research WorldWide Telescope by zooming in from the universe to the earth.


Sensors, Data Acquisition, and Science Narratives

This demo shows several field sensors in the context of research sites, transportation, logistics ,and data capture machinery. The purpose is to show how scientific narratives result from the (often arduous) process of getting sensing equipment operational in the field, and how these narratives, in turn, relate to the modeling community.


SciScope: Web 2.0 Meets Geoscientific Data

For interdisciplinary scientists, engineers, and environmental citizens, SciScope provides an easy way to discover and retrieve environmental quality data as an access point to multiple data repositories offering near real-time and historic data from in-situ sensors as well as bottle samples. In addition to observation results, through SciScope users can access relevant fact sheets to find out about physical properties, toxicity information, and regulations about chemicals or disease symptoms, and taxonomy of pathogenic organisms. Users can also contribute to the system by providing comments on existing datasets or submitting new datasets, hence taking advantage of SciScope to share their data with a larger audience that includes fellow citizen scientists/enthusiasts and the general public.


Tools and Services for Data Intensive Research

We discuss and demonstrate the academic release of Dryad on HPCS and provide electronic copies of the programming guides, algorithms running in Dryad, and hands-on demos that illustrate the simplicity and power of Dryad. In addition we demonstrate services and data on the Microsoft cloud computing platform, Azure. In particular, we demonstrate PhyloD, which was developed in our eScience group as an Azure service for the research community to use.


India Digital Heritage

Advances in the fields of computer vision, graphics, and interactive media make it possible for us to both record and highlight the rich architectural, historical, and cultural heritage that is associated with monuments—in India and across the world. The India Digital Heritage Project is a collaborative initiative between Microsoft Research India, the Government of India, and academia, with the aim of using novel techniques to capture and present various aspects of India’s diverse heritage efficiently, while advancing the state-of-the art in related research areas. The India Digital Heritage demo presents one possible approach to weave a narrative that encompasses the tangible and intangible aspects of heritage.


Microsoft Research Asia eHeritage: Past is Future

Each nation's heritage represents its past glory, its present identity, and its inspiration for the future. However, many valuable treasures that form part of our cultural heritage have been damaged or destroyed by time, weathering, natural disasters, and disasters caused by humans. With advances in modern computer science and technology, we are now in a position to capture the glory of national heritages for present and future generations.
The eHeritage project proposes to apply the latest computing technologies to digital preservation, virtual restoration, and interactive display of the Asia-Pacific region's unique historical treasures. Through this project, Microsoft Research will demonstrate not only how Microsoft advanced technologies and tools can help preserve the heritage of the nations of the world, but also how they can establish a platform for academia, domain experts, and Microsoft researchers to collaborate for the good of society.
The gigapixel digital camera and appearance manifolds technology developed by Microsoft Research and universities are key technologies that can be applied by museums to capture high quality digital data and display. Information about more Microsoft Research and academia collaboration projects can be found on the Web site.


Tool Kit for Visualizing Large-Scale Data

Our tool kit provides a set of Microsoft Silverlight/AJAX controls for visualizing large-scale structured data from various data sources. The controls can be used to expose graphically the structure of the data, trends, and relationships of data properties. We also provide a platform that enables rapid development of a large-scale data-explorer, analysis, and reporting tools. We demonstrate several individual controls and a demo application built atop our tool kit. This demo application lets users explore a large set of article, image, and video data directly and easily.

Renlifang: Web-Scale Entity Summarization

Currently, we manually collect and understand Web information about a real-world entity through search engines. However, the information about a single entity might appear in thousands of Web pages. Even if a search engine could find all the relevant Web pages about an entity, the user would need to sift through all the pages to get a complete view of the entity. This demo presents Renlifang, a Web-scale entity-summarization system that efficiently generates summaries of Web entities from billions of crawled Web pages. Specifically, Renlifang automatically generates:

  • A biography page for a person.
  • A social-network graph for a person.
  • A shortest-relationship path between two people.
  • All titles of a person that are found on the Web.
  • All the structured data we have in our local database about a person.


Swiss Experiment

Situated Interation


Devices for Health Care


Low-Power Processors in the Data Center

This demonstration shows an experimental prototype to study the use of low-power processors in the data center. These processors offer substantial fractions (33 percent to 50 percent) of the performance of the high-performance processors used in Microsoft data centers but consume a disproportionally smaller amount of power (5 percent to 10 percent). Power consumption accounts for as much as 30 percent of the total operating costs of a data center. Across Microsoft Internet properties, about 10 percent of data center CPU cycles perform useful work. Remaining systems run at near full power, because servers and Windows Server do not yet support low-power modes. To study the potential of low-power processors in the data center the Data Center Futures team built a prototype containing 100 dual-core Atom processors. Half are attached to standard hard disks and half are attached to low-power flash storage, to study the tradeoffs of this technology in the data center. The processors are connected through the Monsoon network. This prototype will be used to test new technologies (including flash, optical networks, and FPGA accelerators) and for a variety of experiments, such as powering down idle processors. We believe that data centers in the future will include many more low-power processors. This prototype is the first step in demonstrating their potential cost savings and in developing the algorithms and software to take advantage of these processors.

Closed-Loop Control Systems for the Data Center
This demonstration shows a closed-loop, adaptive control system for Windows Live Search that aims to minimize energy usage while guaranteeing a service-level agreement (SLA) for search response time. Power is a central issue in the design and management of data centers. Power consumption accounts for as much as 30 percent of a data center's operating costs. Idle machines consume a good fraction of this power; only about 7.5 percent of CPU cycles executed in Microsoft data centers perform useful work. Minimizing power usage during periods of low workload could save Microsoft money. Applications deployed in the data centers, however, require a strong guarantee of their performance. For example, Live Search requires the response time of at least 99 percent of queries to be less than 300 milliseconds. A key challenge is to minimize power usage while meeting the desired SLAs. To address this challenge, we present an energy-aware prototype built using 100 low-power Atom processors that execute a Live Search benchmark with a scaled-down, 1-GB search index per node. To meet the 300-millisecond response time, we apply machine-learning techniques that model performance as a function of workload and that set power states (idle, sleep, hibernate) across nodes to save energy. Because transitions between different power states incur a latency of 15 to 30 seconds, our prototype provides a predictor module that switches processors to different power states in advance of workload transitions.


Freehand Interactions with the Omni-Projector

This demo presents a combination of a standard projector with a wide-angle lens capable of projecting data onto the entire, 360-degree surrounding environment from a single position. This setup provides an immersive experience similar to the much more expensive existing planetarium projectors or Virtual Reality CAVE projectors, on which all of the surfaces in the room can receive projections. We have added an infrared camera that shares the wide-angle lens with the projector and is capable of detecting a user's hands and tracking freehand gestures in mid-air, without additional gloves or tracking objects. This demo integrates several Microsoft technologies into a stunning presentation: We offer a hemispherical dome in which users can interact with data from Virtual Earth and WorldWide Telescope.


The MIPS-to-Verilog Compiler

The MIPS-to-Verilog (M2V) compiler translates blocks of MIPS machine code into a hardware design captured in Verilog. With the addition of control transfers across basic blocks, the compiler can handle all the “hot” code in an application, including nested loops and recursive functions. In conjunction with the BBTools and the Giano simulator, M2V offers a complete and automated way to create efficient application accelerators. Shortly, we will provide the first (source) release of the compiler, for academic use.

Scheduling Accelerators on NetBSD

This project uses an established operating system to run applications on the eMIPS dynamically extensible processor. We schedule the extension slots (accelerators) in the best possible way, and without any advanced knowledge of the characteristics of the running applications. The accelerator scheduler is a new loadable kernel module that can be replaced/extended by the end user.


Multi-Core eMIPS

The eMIPS MC microprocessor is the first attempt to implement a secure, multiple core, multi-user platform using an extensible instruction set architecture. The eMIPS MC integrates two or more eMIPS extensible ISA cores using both shared memory and message passing models. The eMIPS MC is a research platform to evaluate and explore application scheduling and acceleration on multiple customizable heterogeneous cores. Targets are the XUP board for multiple cores on a single FPGA, and the Berkeley Emulation Engine 3 (BEE3) for multiple cores per FPGA in a multi-FPGA configuration.


Multi-Touch Input Sensing

This demonstration shows a multi-touch surface based upon the principle of frustrated total internal reflection (FTIR). While FTIR is a popular for creating multi-touch interfaces, most approaches use a projector and digital camera, which make the system large and expensive. This prototype uses infrared LEDs and photo transistors placed along the edge of the surface. A series of filters and algorithms is then used to convert the raw data into discrete touches to be used by a computer. The result is a low-cost multi-touch interface that could easily be attached to any existing display.

High-Speed Communication API for FPGAs

We can speed up many applications by several orders of magnitude using field-programmable gate arrays (FPGAs), but the communication between a host PC and the FPGA itself can often present a problem. Not only does setting up this communication generally require laborious custom software and hardware development, it is often a critical performance bottleneck for the system as a whole. This work eliminates the troubles developers have interfacing with FPGAs by providing a reliable and reusable high-performance software/hardware API infrastructure


LED-based Dance Pad

This demonstration shows how light-emitting diodes (LEDs) can be used to create a reliable user interface for music arcade games like Dance Dance Revolution, In the Groove, and StepMania. One problem many dance pads have is that they have moving parts that eventually wear out. As the pads age, they become less and less reliable, to the point where they are no longer usable to play the game. Users become frustrated by the intermittent failure of the machine and seek other venues, and owners of the machine must repair or replace the broken dance pad. This prototype seeks to avoid this issue by using no moving parts. Instead, the reflectance of visible light off the foot of the user is used to determine whether the user is depressing a particular area on the dance pad. While there is room for improvements with respect to strength and reliability, this prototype already shows a simple, responsive, and effective solution to the "moving parts" problem while adding dazzling lighting effects.


Concurrency Analysis Platform and Tools

Concurrency bugs are difficult to find and reproduce. We demo the Concurrency Analysis Platform (CAP), which provides predictable control over thread interleavings. When a concurrency bug is found, CAP can drive the program along the erroneous interleaving, providing an instantaneous repro. CAP enables multiple concurrency-analysis tools that will be useful for developers and testers. The demo includes CHESS, a systematic unit-testing tool for concurrency; Cuzz, a concurrency fuzzer for obtaining more coverage from existing stress tests; FeatherLite, a lightweight data-race detector; and Sober, a tool for finding memory-model errors.


Solver Foundation: Mathematical Optimization

Microsoft Solver Foundation is a new framework and managed-code runtime for mathematical programming, modeling, and optimization, with a focused goal of helping businesses make near-optimal, strategic decisions. The possible applications cover a vast range: real-time supply-chain optimization, data-center energy-profile management, online-advertizing profit maximization, logistics of large conference scheduling, and risk analysis of investment portfolios. There are also direct applications to graphics and machine learning for which Solver Foundation acts as a runtime for such systems. All of these decisions are encoded through a declarative model specification, one that focuses the modeler and developer of stating the "what" rather than the "how" of the business decision to be made. This rapidly accelerates solution engineering and increases the degree of "what-if?" analysis possible. Solver Foundation has several specific solvers that are good for one or more domain and modeling situations such as linear programming (simplex and interior-point-methods-based), SAT solving, CSP, and quadratic programming. Eventually, solvers will include constrained, convex, non-linear programming.


Audio Spatialization and AEC for Teleconferencing

In multi-party conferencing, one hears voices of more than one remote participant. Current commercial systems mix them into a single mono audio stream, and thus, all voices of remote participants appear to come from the same location. This is in sharp contrast to what happens in real life, in which each voice has a distinct location. We demonstrate technologies to enhance the user experience in multiparty conferencing by using highly realistic, immersive spatial audio for both loudspeakers and headphones. This is proven to improve the conferencing experience significantly, because each participant can easily differentiate the current remote talker and focus on the content being discussed. Multichannel Acoustical Echo Cancellation (AEC) is a crucial component to enable a quality audio experience during conferencing, especially without a headset. We also demonstrate the AEC capability in real time so that the remote side hears only the near-end participant's speech without their own echoes. Performance comparison among multiple AEC algorithms is provided.


Helping Writers Find the Right Words

Writers often need help in choosing words. They might be seeking to introduce variety into their prose or to avoid an awkward or inappropriate phrase. Often, they will consult an online thesaurus. In other instances, as when writing technical documents, writers might need to use terminology aligned with organizational standards and refer to a style manual, possibly stored on a corporate intranet. Conventional thesauri and terminology lists though, are static and usually quite unhelpful with regard to usage in context. We demonstrate a tool that provides writers with inline, contextual thesaurus help and offers a potential path to a new generation of online writing-assistance applications. We combine a paraphrase model, derived from aligned translation corpora and other corpora-based word-similarity data, with a large language model to provide suggested rephrasings that might be appropriate in the writer's intended context. Optional Web-search functionality provides further examples of real-world use. By swapping in, or combining in, new, domain-specific language models, it also becomes possible to modify editing assistance to accommodate specific domains or corporate clients. This application can help both native speakers of English and those for whom English is a second or foreign language.


Commute UX: Dialog System for In-Car Infotainment

After deploying Blue&Me for Fiat and Sync for Ford, in-car dialog systems are morphing from cool gadgets that amaze people and sell more cars to integral parts of in-car infotainment. This raises the bar for the functionality, usability, and reliability of these systems. The presented in-car infotainment system contains novel technologies from Microsoft Research that enable natural-language input; expose a multimodal user interface including speech, a graphical user interface (GUI), touch, and buttons; and use state-of-the-art sound-capture and processing technologies for improved speech recognition and sound quality.



Social Desktop

Today, it is easy to share a Web page or a blog post, because items on the Web have unique IDs: URLs. We do not have this on the computer desktop. Social Desktop adds URLs to the files and folders on your desktop, letting you share anything on your computer with anyone who can click a URL. Persons receiving links can either access via e-mail or comment, tag, and search across all shared items via our Web page. We implement this by using a .NET service, but it is possible to create a universal namespace for every device and data source for a user, providing a universally addressable namespace with:

  • Universal access. The same URL works from any device in the world.
  • Universal sharing.
  • Universal tagging and commenting.
  • Freedom from legacy paths. Data is not limited by file-system concepts. You can have a URL drill into a sub portion of a document or a Microsoft Office PowerPoint deck, or data could come from a Web service or a database.

Social Desktop is a local service that maps the user's local data into a .NET service bus service, enabling local data to be accessible through firewalls. Social Desktop also provides a Web-service view over the same data, with inherent RSS event streams for any container. New data sources can be mapped into the URL hierarchy, enabling a distributed view to be built. There are simple sharing paradigms that enable URLs to be shared temporarily or permanently.



Categorizing the Web

In collaboration with Windows Live Search, we have developed classification technology that assigns one or more topical categories to each Web page as it is crawled. For example, as the page is crawled and indexed, category labels such as "computers," "computers/computer_science," "science," and "reference" are assigned to the page. These category labels provide a higher-level representation of Web pages than traditional content analysis and can be used in many different ways to improve search. We demonstrate several uses of these category labels: annotating search results with relevant categories, enabling users to filter search results by category, increasing the diversity of search results, categorizing users' queries, and understanding the quality of search results.


Understanding Change on the Web

The Web is an ever-changing, dynamiccollection of information. Yet the tools we use to view and search for this information (browsers and search engines) focus on a single snapshot of that information. We present analyses of how Web content changes over time, how people revisit Web pages over time, and how revisitation patterns are influenced by user intent and changes in content. We have developed prototypes that support people in understanding how information with which they interact changes over time. Diff-IE highlights how a Web page has changed since you last visited it and enables you to see how the page has evolved. We also describe a new search model that represents and uses features about the temporal evolution of Web pages to improve ranking and to inform crawl policy.


“I Know It When I See It Search”

Learn about an interactive recommender system that provides an alternative way to find information on a computer. It observes the user’s selections from a controlled set to learn preferences and refine the displayed results by using simple machine learning techniques that are specifically designed to aid in information foraging activities.

Hands-on Math

Hands-on Math will allow people to write and interact with mathematical expressions by freely interchanging direct stylus and multi-touch interactions.

Surface Product Roadmap

This demonstrates key new features of Microsoft Surface’s recent release (SP1), and plots out the challenges and roadmap for the architecture of its user experience and toolkit over future releases. Microsoft Surface is a multi-touch tabletop with an extensible software platform for WPF and XNA development.


MSDN Academic Alliance and Imagine Cup


DemoFest Posters




Integrating Galaxy Zoo and Worldwide Telescope
P1-2 WWT-based MSR and NAOC Collaboration
P2-1 The WorldWide Telescope as a Novel Research and Publication Platform
P2-2 Building a Better Scientist: Computational Science Education
P3-1 From Alaska to Brazil to Antarctica: Environmental science data systems built on Microsoft technology
P3-2 Data Interoperability and Cross Domain Modeling
P4-2 Search by Color
P4-3 Mining Time-dependent Attractive Areas and Movement Patterns from Taxi Trajectory Data
P5-1 Tools for Researchers
P5-2 Chem4Word
P6-1 RAPT, M.U.P.P.E.T.S., and the Games for Learning Institute
P6-2 Game: Themed CS Education: Empowering the Faculty
P6-3 IPRE – Institute for Personal Robots in Education
P6-4 Socially Guided Robot Learning
P7-1 Publishing and Searching Sensor Streams on the Web
P7-2 Towards a Friendlier mPlatform
P8-1 PoolView: Privacy in a Mobile Sensor Web
P8-2 Swiss Experiment
P9-1 Smart Phone-Compatible USB Ultrasound Probe
P9-2 Structured Multi-modal Guidelines on Cell Phones and
Mobile Devices
P9-3 Application of Smart Phone in “Better Border Healthcare Program” (BBHP)
P9-4 An Online Community for Teachers to Support, Observe, Collect, and Evaluate Assisted Communication with Children with Autistic Spectrum Disorders who Use Smartphones as Communication Devices
P10-1 A Caribbean-wide Healthcare Management System
Based on Cellular Phone Technology
P10-2 Random Forest Classification for Automatic Delineation of Myocardium in Real-Time 3D Echocardiography
P10-3 Lowering the Barriers to Cancer Imaging
P10-4 Oncological Image Analysis
P11-1 Snackbot: A Service Robot
P11-2 Human-Robot-Human Interface for an Autonomous Vehicle in Challenging Environments
P11-3 Personal Digital Interfaces for Intelligent Wheelchairs
P11-4 Human-Robot Interaction to Monitor Climate Change via Networked Robotic Observatories
P12-1 FaceBots: Robots Utilizing and Publishing Social Information in FaceBook
P12-2 Multi-Touch Human-Robot Interaction for Disaster Response
P12-3 Survivor Buddy: A Web-Enabled Robot as a Social Medium for Trapped Victims
P12-4 Prosody Recognition for Human-Robot Interaction
P13-1 Microsoft QUT eResearch Centre: Bioinformatics
P13-2 Accelerating Cancer Research using Semantics-driven Technology
P13-3 Archangel
P14-1 CoSBi Lab
P14-2 Development of Systems Biology.Net (SB.NET)
P14-3 .NET-based Clients and Services in the Cancer Bioinformatics Grid (caBIG)
P14-4 Survey Propagation Algorithms in Computational Biology
P15-1 Amyotrophic Lateral Sclerosis (ALS) or Lou Gehrig’s Disease: Searching the Human Genome for Clues Using Chip and Laser Technologies and Computational Biology
P15-2 Shared Genomics: Accessible High Performance Computing for Genomic Medical Research
P15-3 Software System for Genetic Linkage Analysis of SNP Data
P15-4 Direct Brain-Computer Interaction and Functional Neuromonitoring
P16-1 Dynameomics
P16-2 Phenotypic Pipeline for Genome-wide Association Studies
P16-3 PGRx: An Interactive Software System for Integrating Clinical Genotyping with Prescription Drug Safety Assurance
P16-4 An Azure Science Cloud for Drug Discovery
P17-1 Visualizing Voice
P17-2 PowerNet: A Magnifying Glass for Computing System Energy
P17-3 IMPER: A Model of IMagery and PERception
P17-4 Infinite Nonnegative Matrix Factorization
P18-1 Multi-scale Simulations of Soft Elasticity of the Stem Cell and Its Contact/Focal Adhesion with Extra-cellular Environments
P18-2 Cryptography for Pervasive Communications
P18-3 Fast DBMS-Resident Machine Learning and Statistics: Extending the Sloan Digital Sky Survey SQL Server Database
P18-4 Direct Brain-Computer Interaction and Functional Neuromonitoring
P19-1 The Center for Computational Thinking
P20-1 Bio.NET
P21-1 PGRx: An Interactive Software System for Integrating Clinical Genotyping with Prescription Drug Safety Assurance
P21-2 Use Smart Phones to promote Diabetes Self-Management: Robust Elderly in Urban and Rural China
P22 Data Quality Management for Enabling Meta-Analysis in GWAS
P23 System Architecture to Accelerate Database Transactions
P24 Ad-Hoc Floating Point