(v) Innovations

Microsoft Research is staffed by scientists who are first and foremost innovators tasked with a job they expect to never complete. Their assignment: to continually advance the state of the art in computer science. Given the freedom to pursue the subjects they are passionate about, they look for gaps in current technologies and dedicate themselves to filling them. They envision what comes next, without worrying about product-delivery deadlines. When their ideas develop into innovations with real-world applications, the researchers work with a team of technology-transfer agents or directly with product teams to add these innovations into Microsoft software and other offerings.

Nearly every Microsoft product in the marketplace today has been influenced by Microsoft Research, including Windows Server™ 2003, Office 2003, Windows XP, Office XP, the Xbox gaming system, MSN 8, Microsoft SQL Server, Windows Media® Player 9 Series, and Smart Personal Objects Technology. Occasionally, researchers involved with a project even transfer to the product development group where their initial ideas will take shape.

Breakthroughs, Large and Small

Although some people believe that being innovative or making a breakthrough means “getting there first,” Microsoft Research takes a much broader view. A technology breakthrough or innovation can take years to evolve, and its impact may not be fully realized until years later — as in the case of Internet technology or even e-mail. Microsoft believes that a combination of invention and popularization makes a technology or product innovative.

No matter how far-reaching or abstract their projects might be, Microsoft researchers consistently strive for results that eventually will solve real-world technology challenges. In many cases, their results are shared and applied across multiple research areas. The following are examples of the challenges Microsoft Research is dedicated to solving and some of the current projects that target these challenges.

Making Computer Interaction Easier, More Natural

It would be much easier to interact with technology if people could speak directly to a computer and write directly on the screen instead of being limited to the technical language and interfaces of PCs and handheld devices. Microsoft researchers are enabling computers to more accurately interpret their surroundings and, as a result, better understand what the user intends to do.

Examples of this research include:

Location awareness. Microsoft Research is developing new ways to do everything from locating nearby friends who are out of sight (but not out of cellular phone range) to improving remote learning on college campuses. The Social Computing Group at Microsoft Research is integrating desktop and mobile devices to create prototype services such as “NearMe,” which notifies a user when his buddies are nearby—using cellular network information, a cell phone’s contact list, and short message service messaging. Another prototype service, code-named “Project X,” seeks to improve learning among groups of college students who aren’t always able to study together in person. In collaboration with the MIT architecture school’s iCampus project, “Project X” uses technology developed by Microsoft researcher Victor Bahl to provide students with increased awareness of other students’ physical location and presence.

Using Computer Science to Combat HIV/AIDS

The Challenge

According to the Global Health Council, nearly 30 million people worldwide have died of Acquired Immunodeficiency Syndrome (AIDS), and the disease claims another victim every 10 seconds. Approximately 5 million people are infected each year, and an estimated 40 million people are living with human immunodeficiency virus (HIV), the virus that causes AIDS.

Despite years of vigilant effort to develop an effective, long-term vaccine for HIV, medical scientists remain stymied. Because HIV mutates each time it infects a new person, finding patterns in the mutations is the key to finding an effective long-term vaccination. To find the patterns requires analyzing huge amounts of data.

The Solution

Microsoft researchers David Heckerman and Nebojsa Jojic are applying software algorithms similar to those used to manage computer databases, compress digital files, or block spam e-mail to overcome roadblocks in the hunt for an HIV vaccine.

They have developed software that will detect patterns much more quickly than medical scientists can. In just a few months, the software completed the same analysis that had taken medical scientists more than five years. Microsoft’s innovative work, which is being done in collaboration with AIDS researchers at the University of Washington in Seattle and the Centre for Clinical Immunology and Biomedical Statistics at the Royal Perth Hospital in Perth, Australia, is helping biomedical researchers discover the right set of genetic components for a vaccine to fight the virus.

Natural language processing and speech recognition. By designing computer systems that can analyze, understand, and generate languages that humans use naturally, Microsoft Research aims to enable people to talk with computers as though communicating with another person.

Machine learning. Microsoft Research is seeking ways for computers to better understand users and tasks and do a much better job of managing information, computing resources, assistance, and user attention.

Vision technology. Projects that range from image-based rendering and animation to human tracking and 3-D scene reconstruction are influencing the development of the next generation of user interfaces.

Telepresence. This research focuses on enabling people to feel as if they are present at an event while physically in another place or time, through the use of digital media such as video, audio, images, and animation.

Improving Software Development

Performance, quality, manageability, and productivity are crucial to developers when creating software in today’s increasingly complex and competitive market. Microsoft Research is inventing new programming tools, methodologies, and techniques to help developers build better software and deliver this software to customers faster and at lower prices. Key projects supporting research in this area include:

Development tools. This research is aimed at advancing the design, development, debugging, and testing of software, so programmers can be more creative, imaginative, and productive.

Programming principles and tools. Formal techniques and models are being developed for understanding programs, programming abstractions, and languages.

Software engineering. Improving the methods, notation, and tool support for high-level system design and analysis is the goal of this research.

Improving How Systems Store, Retrieve, and Present Information

At work and at home, people rely on an ever-growing sea of data from the Internet, e-mail, business transactions, and other sources. As computers store greater volumes of this data, they need to offer increasingly intelligent tools for protecting, extracting, and analyzing this data. From there, users also expect the systems to present the data faster, more intuitively, and in a wider variety of formats. Computer scientists at Microsoft Research are exploring ways to design operating systems, architectures, and components that can meet these demands as well as those of the future. Microsoft Research projects include:

Search. Microsoft Research is creating search tools that do more than provide long lists of documents. Technologies now under development enable people to more effectively navigate results and “train” the search engine to provide the desired information faster. Other tools provide actual answers to questions, rather than lists and links.

The Web Search & Data Mining Group aims to help improve how people analyze, organize, retrieve, and visualize information. Although most Web search methods essentially only rank and retrieve data from documents, this group uses methods that delve much deeper, searching at the object level to capture increased knowledge and intelligence about data.

Other Microsoft researchers are developing search algorithms that provide results based on more than the query a person types into the computer. They are pursuing techniques that also factor in a person’s interests, based on previous search queries and Web pages visited, along with documents and e-mail the person has read or created. This information is then used to re-rank Web search results.

Peer-to-Peer (P2P) file sharing. Microsoft researchers are developing ways to make online exchange of large files among numerous people more efficient and useful. A P2P coding system developed by Microsoft researchers proved to be 20 percent more efficient than another leading P2P technology in tests conducted by the researchers. Code-named “Avalanche,” the model also ensures content providers are uniquely identifiable and helps prevent unauthorized parties from offering content for download. Another prototype technology developed by Microsoft researchers in Asia may eventually enable organizations to more cost-efficiently store and manage petabytes of information. The technology uses P2P for self-managing and self-healing of information, particularly large volumes of seldom-changing information. It is designed to use massively parallel systems to repair faults within the information, reducing the vulnerability of data loss.

Mobility. To integrate computing throughout their lives, people need more and better ways to communicate and control digital information when they are away from home or work. Several groups within Microsoft Research are developing these mobile tools and technologies. The Socio-Digital Systems Group combines psychology, sociology, computer science, and hardware engineering to design technology capable of supporting people throughout their lives. The Sensors and Devices Group is exploring how sensors, embedded devices, and other systems can make computing more mobile. The group’s projects include “SenseCam,” a badge-like “black box” data recorder that automatically captures images and sensor data as people move around during the day, creating a visual blog that can then be played back at high speed to share particular experiences with friends or as a way of looking back over a day. Another project, code-named “HomeNote,” provides a way to communicate to a specific place rather than a person, via a touch-sensitive flat-panel display that can receive and display communications such as e-mail, text, and picture messages.

Security/privacy. Microsoft research projects aim to protect customers and build and maintain their trust in computing. A Microsoft Research tool called “Strider GhostBuster” combats rootkits, which is malicious software that attempts to surreptitiously take control of a computer, while a similar tool that Microsoft Research developed, Strider Honey- Monkey, detects programs that use Web servers to exploit browser vulnerabilities so malware can be installed. Strider HoneyMonkey helps the Microsoft Security Response Center react to threats in a timely way and also enables investigators to detect and analyze Web sites hosting malicious code.

Cryptography and anti-piracy. Projects include researching new encoding methods and applications to enhance privacy and security, working with standards bodies to develop security protocols, and providing internal security consulting on Microsoft products.

Exploring and Solving Tomorrow’s Most Complex Computing Problems

Even a casual glance at the progress of computing over the past 30 years — from early mainframe systems that filled an entire room to today’s PCs that fit in a shirt pocket — shows how rapidly the boundaries of technology are expanding. But many significant challenges remain. Microsoft Research tackles some of the toughest problems in computer technology through its efforts in the following areas:

Algorithms and other mathematical methods. Microsoft Research is devising novel formulas and procedures to advance the state of the art in such areas as natural language processing, speech recognition, handwriting recognition, and time-series prediction.

Theory. Researchers apply the principles of mathematics, statistical physics, theoretical computer science, and other disciplines to probe the limits of computational speed, problem-solving, and decision-making capabilities.

Building a Smarter Spam Trap

The Challenge

For years, consumers and businesses had to rely on one-size-fits-all software tools and services to slow the flood of spam into their e-mail inboxes. However, spammers increased the volume of their unwanted commercial correspondence and developed ways of fooling most spam filters. The constant deluge has reduced consumer trust in e-mail, poses a security risk for children, burdens computer networks, and wastes time, money, and resources of e-mail users around the world.

The Solution

Microsoft Research and the Safety Technology and Strategy Group developed SmartScreen™–Exchange Intelligent Message Filter, which, following its release in 2003, enabled computer users to block more than 95 percent of incoming spam allowed — a total of 3 billion pieces of junk e-mail a day. With the Exchange Intelligent Message Filter, customers can score each incoming e-mail message to help filter spam before it reaches their inbox, making spam easier to detect and more difficult to profit from. SmartScreen Technology helps maintain e-mail as a viable and valuable method of communication for the hundreds of millions of customers who use MSN, Hotmail®, Office, and Exchange. 

Next >

Contents

• Advancing the Frontiers of Computing
• Microsoft Research: Making the Impossible Possible
• Creating a Foundation for Technology Breakthroughs
• Assembling an All-Star Team of Research Talent
• Imagining What Comes Next
• Putting Innovation to Work in Microsoft Products
• Advancing the State of the Art
• Planting the Seeds of Future Innovation
• Building a Global Think Tank