A journey through the corridors of one of India's best research centres provided not only interesting insights about the research scenario in the country, but also about what makes MSR India a dream destination for any researcher.

"A scientist in his laboratory is not a mere technician; he is also a child confronting natural phenomena that impress him as though they were fairy tales," said Marie Curie, one of the greatest scientists and researchers of the 20th century.

How well the words reflect what makes a researcher. Some equally enchanting thoughts on the passion that possesses and stimulates a researcher were shared by P. Anandan, MD, Microsoft Research (MSR) India, who, interestingly, describes researchers as "the artists of science." However, there is a lot to a researcher and a research centre, besides this philosophical outlook. There is a method in the way a research centre gathers like-minded researchers to work on and explore unknown and unforeseen vistas.

MSR India is one such research facility, which in a short span of three years, has evolved as an epicentre for ground-breaking research work; where great

minds converge and work consistently on projects aimed at advancing the state-of-the-art in different research areas. The lab at Bangalore joined Microsoft's band of research facilities in 2005. Anandan, who, along with a team of researchers, set up the Indian arm of MSR, dreamt of seeing the facility emerge as a world-class global research centre. And to his happiness and pride, in a short span of three years, the lab stands shoulder to shoulder with other MSR facilities the world over. As Anandan reveals, "MSR India has to its credit more than 120 published research papers; we have won Best Papers awards in several major conferences worldwide. This is an important metric for us to evaluate our progress as the papers we publish are reviewed by some of the top researchers in the field before being accepted in conferences."

While at present, MSR worldwide has over a thousand researchers who are working in 60 different research areas, the Indian arm has a team of over 44 researchers working on seven core research areas, with several innovative projects.

Why India?

It seemed worthwhile to probe into the factors that drove MS Inc to set up a lab in India. Was it talent, the growing economy, or a plan to identify the needs of the huge consumer base in India and devise products to cater to them?

Anandan responds, "A research lab is not like a product group, where you go and hire people to do a certain task. Instead, it is a place where you gather researchers and ask them what they want to do. So people are more important than anything else. A lot of people think that MS sets up a lab in a certain country because there is a certain market that it wants to capture, or a certain problem that it wants to solve. But that is not true. MSR always chooses a location/country on the basis of the most important factor, which is talent."

Indian talent has been acclaimed for years for its brilliance, and this was the key reason for MSR to zero in on India as a location for one of its research units. "In India, we don't have many PhDs, but we have a tremendous [number] of B-Techs and M-Techs who are really fantastic," avers Anandan.

Considering the progressive trend that the economy is following, an emerging reality is that more and more scientists of Indian origin the world over are contemplating a return to their homeland, and not many young engineering graduates are keen to leave for distant lands to pursue careers. Even Anandan, who was the head of the Interactive Visual Media group at the MSR Redmond lab, returned to India to set up this research facility. He observes, "It's no longer a case where we have to moan any more about losing our best talent to the world."

Also, there is a shift of attitude that's happening, which is also drawing more and more firms to set up their research facilities in the country. Anandan observes that years ago, when the IITs were set up, their primary goal was to be great teaching institutions, in order to produce the best engineers for the country. Over time, professors in these institutes have become more engaged in research. Today's faculty members are keen to become world-class researchers, in addition to being good teachers. MSR India enjoys a healthy interaction with Indian academia for research projects, and also to support their growth. He adds, "Besides, India represents a large portion of the world that is an emerging/developing market. There are problems related to emerging markets that we may be able to think about a lot better by being here in India. Those are some of the primary factors that are behind setting up the lab in India."

To research is to take a global view

So what is the agenda or the primary goal of MSR India's operations? Anandan reveals, "The first MSR lab was established in 1991, at Redmond, as a basic research unit for MS Inc. For the Indian lab too, the model is the same as that of the other labs: that of being an independent entity and research unit, where researchers can do things that they want. There are three mission goals that are true for all MSR labs worldwide, as per Anandan:

• To do the best research possible, to advance the state-of-the-art in different research areas worldwide--not only for MS, but for the advancement of science.
• To impact MS products, and help/work with product groups to improve MS products, through research and technologies. Every MS product has some component that is a contribution from MSR labs.
• To look toward the future for new possibilities. Product groups have a timeframe, but research groups take a long-term view, and identify technologies that will define the future.

Opportunities, unlimited

Anandan feels that being part of MS is both an opportunity as well as an obligation: an obligation to advance the state of technology for the benefit of computer users everywhere, but more importantly, an opportunity to learn, through the interaction with product groups, what end users really need and want.

He says, "It's a huge opportunity if our research is made use of by billions of people, compared to a scenario where we publish a paper that gets applauded in a conference hall." MSR also organises a yearly event called TechFest, where researchers from MSR labs across the world stage demos of their projects. Thousands of people from the product groups walk through and interact with researchers to understand their projects; if they like some idea, they take a project or a part of it to productise. However, it is one thing to come up with innovative project findings, and quite another to transform the research breakthrough into a product. As Anandan explains, "If it takes two years to solve a research problem, then it could take four more years to productise it. Also, not everything that we work on is productised. Besides, our code can be sloppy, but the product group has to make sure that [theirs] is usable, maintainable, and re-usable."

Core research areas

MSR aims to continue expanding its research team and operations in 2008. The group will pursue a research agenda built around seven core areas that are central to MS's long-term vision and strategy:

• Algorithms: This research group focuses on 21st-century algorithms that seek to address new demands in today's environment of computing with massive data. Areas such as streaming algorithms, directed data sampling, machine learning, approximation algorithms, and integration of numerical and combinatorial methods are expected to play an important role in the years to come.
• Cryptography, security and applied mathematics: Focuses on applied mathematics, multimedia security, cryptanalysis, cryptographic protocols and primitives, and also studies various aspects of security related to computer systems and mobile devices.
• Digital geographics: Involves representation, analysis and visualisation of digitised geographic information, and combines work from relevant fields of computer science, such as GIS (geographic information system), graphics, user interfaces, spatial databases, image processing and computer vision.

An interesting project in this group is the 'Robust Location Search Engine', inspired by the scenario in India. Anandan explains that usually searching for location queries requires inputs in a specific format. Unfortunately, in India, we don't have US-style addresses; we have details like: 'near the police station', 'adjacent to Ganesha temple', etc. The problem is that if you give such queries to current map services, they are unable to find a location. Thus, MSR India decided to develop a new address search engine to accept terms like those above, based on intersecting locations, and independent of any address format, and even robust to spelling errors. Having already demonstrated it, the project team is working with the Windows Live team to incorporate it in their product suite.

• Mobility, networks and systems: This group conducts research in areas such as Internet-scale systems, distributed systems, network protocols, wireless networking, mobile computing and sensor systems, seeking to advance the state-of-the-art in these areas through measurement, design and prototyping of real-world and experimental networked systems.Recently, this group did a small project called Combine, demonstrating how a group of friends with cell phones can download a huge media file by automatically pairing with other phones, getting all the phones to share the download (split into different chunks), and then re-build and share the download by combining the bits together over Bluetooth or Wi-Fi. Anandan says that they realised that this was something that will become a trend - a collaborative download mechanism to pool together WWAN (wireless wide area network) links, using high-speed local wireless links as the glue. Their work focuses on a number of challenges, including incentives for collaboration, fast and energy-efficient collaboration group formation, protocol mechanisms to enable deployment with little or no modification to the existing Internet infrastructure, and security and privacy.
• Multilingual systems: This focuses on developing true natural-language-neutral approaches in all aspects of linguistic computing, and on fostering and participating in a healthy research ecosystem in Indian languages computational linguistics. An interesting project is Linguistic Tools in Indian Languages, aimed at developing basic tools such as morphological analysers, parts-of-speech taggers, parsers, named-entity recognisers, etc, in Indian languages.
• Rigorous software engineering: The group studies software engineering issues in developing large-scale software systems, to come up with new tools, languages, and methodologies that dramatically increase the productivity of software development. The focus is on code understanding, configuration tools and design tools.
• Technology for emerging markets: This group, led by Kentaro Toyama, assistant managing director, MSR India, combines technical and social science research to address the needs of the people of emerging economies. It seeks to address both those who are increasingly consuming more technology services and products, and those for whom computing technology remains out of reach. Anandan states that the aim is to explore different models of technology innovations to work out the best options to make ICT useful in the context of socio-economic development.

Two very interesting projects undertaken by the group include:

Text-free user interfaces: These are design guidelines for computer-human interfaces that allow any non-literate person, on first contact with a PC, to immediately realise useful interaction with minimal or no assistance. Anandan informs us, "Rigorous user evaluations show that text-free designs are strongly preferred over standard text-based interfaces; first-time, non-literate users are, in fact, able to navigate through text-free UIs meaningfully."

Split-screen UIs for small businesses: Often, in rural areas, due to economic constraints, there are fewer computers than potential users. To try to solve that problem for schools, MSR India had worked earlier on MultiPoint, which enables multiple children to use a computer simultaneously, with each child having his or her own mouse and on-screen cursor. A similar innovation for small businesses that do not have funds for an elaborate IT infrastructure, the Split-screen UI allows users to plug in a second mouse and keyboard, and split the screen into two sections, each running a separate Windows session - allowing two users to use one PC simultaneously and independently. This may not work for all applications, but can work well for data entry and lookup, simple document processing and browsing.

Talent and recruitment

MSR India gets plenty of interest and applications, both from India and abroad - but the standards are set very high. Anandan smiles, "Any researcher in MSR India should be capable of being a researcher in MSR anywhere, worldwide; and any researcher at MSR worldwide should be capable of being a professor at the best university in the world." Setting such high standards helps in the long run: “If you stick to standards, you may be a bit slow in recruiting, you may not get [many people], but then they will be self-motivated, proven to be good, excited and passionate about what they are doing, and won't require to be managed... and this last factor is the most important."

Anandan's grouse is that, "We still don't have enough students going into research. The best IIT graduates and so on, used to earlier go to world-class universities for PhDs but now they are looking to join places like i-Banks. Yet, the truth is that there is a lot more talent; India has a strong reservoir of engineering talent--a world-renowned fact -- but there are only about 35 to 50 PhDs in computer science from Indian universities." In contrast, there are a lot more PhDs of Indian origin in the US than in India in computer science.

To deal with this situation, MSR India has an interesting strategy. Anandan reveals, "We have this position of assistant researcher--a two-year position for young engineering graduates with a passion for research. After that, these researchers can apply in places like MIT, UC Berkley, or any other world-renowned university. We have fifteen to twenty such people; six to seven have already completed their term, and gone off to do their PhDs." However, 57 per cent of the MSR India research staff has completed its PhDs.

The stimulator

MSR India is a campus of 20,000 square feet, with a lively environment that fosters creativity. The walls are glass; people can even write on them, as long as it is erasable. "It's a very open and cooperative environment. People gather in the collaboration rooms, there are always discussions happening. We also have frequent parties in the lab. It's not as quiet a place as many people perceive it to be," smiles Anandan.

A lot seems to be happening at the MSR India labs! In a mere span of three years, the facility has established a strong base, gathered some of the best talent, and is ready with many interesting innovations from its repertoire of research projects. We wish them success in the future!