Technology Research in India

1
Technology Research in India
A Case Study of Microsoft Research India

• Kentaro Toyama, PhD
• Assistant Managing Director
• Microsoft Research India
• Presentation to Technology Management Program
  George Mason University
• May 23, 2007 Bangalore

2
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

3
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

4
India

• People
• 1.1 billion people
• Over half under 25 years old
• 22 languages
• Annual incomes 100-100M
• 28 states
• Area
• 1/3 the area of United States
• Technology
• 20M PCs, installed base
• 140M mobile subscriptions
• 7M each month

Roads in India
Sources CIA Factbook, TRAI, CNN
5
India, a Personal View
My first trip to India (2004)
6
India, a Personal View

• People
• 1.1 billion people
• Over half under 25 years old
• 22 official languages
• Annual incomes 100-100M
• 28 states
• Area
• 1/3 the area of United States
• Technology
• 20M PCs, installed base
• 140M mobile subscriptions
• 7M each month

• but, power held by few
• tremendous energy and optimism
• incredible diversity, EM microcosm
• reminiscent of European Union
• impact of weather (ubiquity of agriculture)
• huge interest in PCs, by everyone
• mobiles, mobiles, everywhere

Huge potential opportunity for Microsoft. But,
there are new challenges that neither India nor
Microsoft have ever faced before.
7
Rural school in Chinhat, Uttar Pradesh
8
Rural village with a VSAT Internet connection
near Bhopal, Madhya Pradesh
9
A small Internet café on a market street in a
town near Bombay
10
Infosys campus, Bangalore
11
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

12
Microsoft Research

• Established 1991
• 700 full-time staff in 5 locations
• Redmond Beijing Cambridge, UK Mountain View,
  CA Bangalore
• Over 60 computer-science research areas
  represented
• Regular publications in major CS journals and
  conferences
• Contributions to Microsoft products
• Ranging from development tools, data mining,
  photo editing, text-to-speech, grammar checking,
  spam filtering, etc.
• http//research.microsoft.com

MSR HQ in Redmond
13
Microsoft Research Mission

• Goals
• World-class academic research
• Impact on Microsoft products and business groups
• Collaborations with external institutions to
  further technology research worldwide

MSR HQ in Redmond
14
MSR India Mission

• Goals
• World-class academic research
• Impact on Microsoft products and business groups
• Collaborations with external institutions to
  further technology research in India, South Asia,
  and Emerging Markets

Microsoft Research India In Sadashivnagar,
Bangalore
15
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

16
Microsoft in India

• Six subsidiaries
• Sales Marketing 1990
• Software Development 1999
• Technical Support 2003
• Consulting Services 2004
• Research 2005
• IT Support 2005

17
MSR India at a Glance

• Established January, 2005
• Six research areas
• Cryptography, Security Algorithms
• Digital Geographics
• Mobility, Networks Systems
• Multilingual Systems
• Rigorous Software Engineering
• Technology for Emerging Markets
• Currently 50 full-time staff
• large internship program
• Collaborations with government, academia,
  industry, and NGOs

Microsoft Research India In Sadashivnagar,
Bangalore
http//research.microsoft.com/india
18
People

• Full-time staff total 49
• Technical staff total 43
• 20 with PhD (46)
• 5 PhD from India
• 15 PhD from abroad
• Location before joining
• India 23 (53)
• Abroad 20 (47)
• 6 women, 34 men (16 women)
• Competition IBM, Yahoo!, Bell Labs, HP Labs,
  Google, Etc.

Group photo (January, 2006)
19
Internships

• Total internships in 2006 81
• To date 122
• Institutions represented (40 total)
• India
• BITS Pilani
• IIIT-Bangalore
• IIIT-Hyderabad
• IISc
• IITs (Delhi, Madras, Bombay)
• ISI Calcutta
• Abroad
• Carnegie Mellon
• UC Berkeley
• University of Washington
• Georgia Tech
• Harvard
• Oxford

Lab size over two years
20
Conferences, Etc.

• Conferences, workshops, and tutorials
  co-sponsored or co-organized by MSR India in
  2006
• Wireless Networking Summit (WiNS) April 2006,
  Goa
• 2 days, 80 participants (Victor Bahl, Uday
  Desai, Mythreyee Ganapathy)
• ICASSP Tutorial on Text-Dependent Speaker
  Recognition May 2006, Toulouse
• 1 day (Amitav Das)
• IEEE/ACM Intl Conf. on ICT and Development
  (ICTD) May 2006, Berkeley
• 2 days, 200 participants (Raj Reddy, Anno
  Saxenian, Kentaro Toyama)
• Cryptography summer school May-Jun 2006,
  Bangalore
• 21 days, 80 participants (Venkie, Vidya
  Natampally, Anandan)
• Afternoon with Design Aug 2006, Bangalore
• 1/2 day, 60 participants (Archana Prasad)
• Virtual Earth Academic Summit Nov-Dec 2006,
  Redmond
• 2 days, 60 participants (Gur Kimchi, Kentaro
  Toyama)
• IJCAI Workshop on AI for ICT and Development Jan
  2007, Hyderabad
• 1 day, 20 participants (Kentaro Toyama, Rajesh
  Veeraraghavan, Krithi Ramamritham, Anupam Basu)
• IJCAI Tutorial on Design in ICT and Development
  Jan 2007, Hyderabad
• 1/2 day, 30 participants (Bernardine Dias, Rahul
  Tongia, Kentaro Toyama)
• Part-of-Speech (POS) Tagset Workshop Jan 2007,
  Bangalore

• MSR sponsorship and co-organization

• MSR researcher led

Key
21
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

22
Cryptography, Security, and Algorithms
Goals
Mathematical and practical aspects of

• Cryptographic primitives
• New paradigms for cryptanalysis protocols
• System and code security
• Algorithms
• Error-correction problems in machine learning

23
Fast Arithmetic for Elliptic Curve Cryptosystems
David Jao S. Ramesh Raju Venkie

• Markov-Chain-based analysis of a number system
  tailored for faster elliptic curve arithmetic in
  cryptographic systems
• History
• Summer 2005 Early explorations
• Early 2006 Refinement and verification
• Fall 2006 Tech transfer
• Transferred
• Arithmetic algorithms
• Whiteboxing tool for digital rights management

Elliptic curve addition
Collaboration with Windows DRM
24
Digital Geographics
Goals

• Invent new technologies to support digital
  mapping and location-based services
• Conduct research in
• Graphics
• User interfaces
• Spatial databases
• Image processing
• Visualization
• Etc.

Auto-generated panoramic map (Neeharika Adabala)
25
Virtual India
Joseph Joy and Virtual India virtual team

• Multilingual online map of Indian cities
  generated from Survey of India data.
• History
• Jan 2005 MoU signed with Ministry of Science
  Technology
• Jan 2006 Online prototype unveiled by Minister
  Kapil Sibal
• Summer 2006 Tech transfer
• Transferred
• Tile generation pipeline
• Transliteration
• Person transferred, also ?
• Udayan Khurana Thapur Institute of Engineering
  and Technology
• (MSR intern ? IDC employee)

Kannada and Hindi views of Bangalore in Virtual
India
Collaboration with Virtual Earth / Windows Live
Local
26
Mobility, Networks Systems
Goals

• To conduct research in networked systems
• Internet-scale systems
• Distributed systems
• Network protocols
• Wireless networking
• Mobile computing
• Sensor networks

COMBINE collaborative downloading
27
Mobility, Networks Systems
Sample Projects

• Proximity Networking
• SPACE Lightweight Peer-to-Peer Trust
• ACM HotNets 2006
• COMBINE Collaborative Downloading
• IEEE HotMobile 2007 (to appear)
• WiFiAds Location-sensitive Advertising
• IEEE HotMobile 2007 (to appear)
• Sensor Networks
• SenSlide Sensor System for Landslide Prediction
• ACM SIGOPS Operating Systems Review, 2007 (to
  appear)

SPACE establishing peer-to-peer trust
28
Multilingual Systems
Goals

• To develop seamless natural-language-neutral
  approaches in all aspects of linguistic computing
• To help create an Indic-language research
  ecosystem

wikiBABEL project
29
Multilingual Systems
Project Overview
30
Rigorous Software Engineering
Vision
Improve productivity by bringing rigor to
software development in the large Look at
Microsoft platform from the point of view of
partners and customers, and conduct research to
improve their productivity
RSE team, summer 2006
31
Netra
Prasad Naldurg Stephan Schwoon Sriram
Rajamani John Lambert

• Analysis tool for finding security flaws in
  access-control configurations
• History
• Fall 2005 Prototype developed
• January 2006 Presented to TAB
• Fall 2006 Tech transfer
• Transferred
• Specification language
• Analysis tool
• Visualization tool

Netra schematic
Collaboration with Secure Windows Initiative
32
Why India?

• Cryptography
• Extremely bright math students
• Digital Geographics
• Strong interest in mapping
• Mobility, Networks, and Systems
• E.g., Fastest growing mobile-phone market
• Multilingual Systems
• 22 national languages, multilingual speakers
• Rigorous Software Engineering
• E.g., worlds most advanced system integrators

33
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

34
Technology for Emerging Markets
Goals

• Social
• Understand (potential) technology users in
  emerging-market countries
• E.g., urban middle-class
• E.g., rural entrepreneurs
• Technical
• Identify applications of computing that support
  socio-economic development of poor communities
  worldwide

Sugarcane co-op member using a mobile phone to
check on details of his harvest in Warana,
Maharashtra
35
Rural Microfinance and IT
Peri-Urban Internet Cafes
Technology for Emerging Markets
Can computers help existing structures for rural
microfinance?
Study of Internet cafes in areas between urban
and rural
Sample Projects
Aishwarya Lakshmi Ratan MPA, International
Development
Nimmi Rangaswamy PhD, Sociology
MultiPoint for Education
Computers in Agriculture
Digital Study Hall
Experiments with computing and communication
systems in agriculture
DVD exchange over postal service and TVs as
display for rural education
Multiple mice to multiply the value of PCs in
schools.
Udai Singh Pawar MSc, Physics
Randy Wang PhD, Computer Science
Rajesh Veeraraghavan MS, Economics and CS
Preventative Healthcare
Participatory Development
IT and Microentrepreneurs
UIs without text for users who are illliterate
and may never have seen a computer before
An analysis of ICT in development projects using
the lens of post-colonial theory.
Information ecology of small businesses in
developing markets
Indrani Medhi MDes, Design
Jonathan Donner PhD, Communications
Savita Bailur PhD cand., Information Sys.
36
Text-Free UI
Indrani Medhi Aman Sagar Kentaro Toyama

• Identify design principles for designing UIs that
  allow non-literate, first-time computers user to
  gain value from their first interaction with a
  computer.
• Group Tech for Emerging Markets
• Title Text-Free User Interfaces for Illiterate
  and Semi-Literate Users
• Authors Indrani Medhi, Aman Sagar, Kentaro
  Toyama
• Venue IEEE/ACM First Intl Conference on
  Information and Communication Technology and
  Develompent, UC Berkeley, May 2006.

Text-free user interface?
Selected for special issue of ITID ICTD2006 Best
Papers!
37
MultiPoint
Udai Singh Pawar Joyojeet Pal Rahul Gupta Kentaro
Toyama
(was MultiMouse)

• Multiple mice cheaply multiply the value of PCs
  in resource-constrained schools.
• History
• Summer 2005 ethnographic studies in rural Indian
  schools
• Fall 2005 First prototype
• 2006 Tech transfer
• Transferred paradigm and SDK
• Dissemination through Imagine Cup 2007

MultiPoint user studies
Collaboration with Market Expansion Group and
Education Core
38
Outline

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

39
What the Press Says (1/3)

• After the tech boom - what's India's next big
  thing?
• Following the dramatic success of India's IT
  services companies over the last decade, many
  industry watchers are now hungrily awaiting the
  country's next trick - to create a software or
  hardware giant along the lines of an Indian
  Google or an Indian Intel.
• Steve Ranger, silicon.com, April, 2007

http//www.silicon.com/research/specialreports/ins
ideindia/0,3800013641,39166051,00.htm
40
What the Press Says (2/3)

• India and Innovation at Davos
• The next round of outsourcing is outsourcing
  innovation. And here India is the center of the
  global economic universe. By language, training,
  education, and diasporadic disposition, India's
  role in the world economic is more brain-driven,
  service-driven and ultimately innovation driven.
  And India, chaotic though it may be, is free and
  democratic. You don't have an army of censors
  watching over the internet and blogs, as you do
  in China.
• Bruce Nussbaum, BusinessWeek (Jan 20, 2006)

http//www.businessweek.com/innovate/NussbaumOnDes
ign/archives/2006/01/india_and_innov.html
41
What the Press Says (3/3)

• India an innovation giant? Yes!
• innovation includes services, manufacturing
  processes, customer facing and back-end process
  in services.
• To me, Bharati Airtel is the most innovative
  company of our times for the way it has created a
  successful business model. The company has
  outsourced everything but its customers, thus
  being able to offer mobile telephony at 10 paise
  a minute nowhere in the world can you get such
  rates.
• S. Kapur, Business Standard, February 21, 2007

http//in.rediff.com/money/2007/feb/21guest.htm
42
Firms with Labs in India
43
Conclusion

• India
• Microsoft Research
• Microsoft Research India
• Overview
• Research Groups
• Technology for Emerging Markets
• Beyond Microsoft

44
Rajkumar Riots

• Kannada film actor, Rajkumar passed away on April
  12, 2006. He lived within several blocks of MSR
  India.
• Fans and riff-raff rioted, imitating riots
  following his kidnapping in 2000.
• Most building windows were broken.
• No physical harm to lab members.
• Building fully restored, thanks to insurance.

45
Code4Bill Contest

• The Prize Write code for Bill Gates, reporting
  to his technical assistant for one year
• Seven-month contest run by MS India DPE
• Three rounds of puzzles and coding challenges
  online
• Two rounds of interviews
• Final round of presentations, winner selected by
  jury
• 24,000 contestants
• 19 in last round, all offered (and took)
  internships with Microsoft.
• Four interned at MSR India.
• And, the winner is
• Abishek Kumarasubramanian
• IIT-Madras
• Earlier worked at MSR India as an intern
• Currently working as an assistant researcher at
  MSR until visa issues clear

Abishek with Bill Gatess then technical
assistant, Alex Gounares
46
Thank you!

• http//research.microsoft.com/india
• Questions? kentoy_at_microsoft.com

47
(No Transcript)