Machine Translation: English to Indian Language

1
Machine TranslationEnglish to Indian Language

• Proposer Peeyush Bajpai
• Name of the company Indicus NetLabs Private
  Limited
• Language/Language pair English to Hindi
• Category of contribution Development and Spread
  of Machine Translation from English to Hindi

2
Strengths
Organizational
Organization
Technical Capabilities
Indicus
Focused Dedicated Team
Past Experience
3
Organization
Indicus Analytics
Indicus NetLabs

• Part of the development process
• Facilitating the process for everyone to gain
• Entrepreneurial in our aiding development
  activities
• Credibility
• Media Coverage
• Client Orientation
• Highly exacting clients in both software and
  research
• Working with Researchers (Academics and Policy
  Makers)

• Focus on Indian Language Technology
• Facilitating Indias Development
• Passion to make a difference

4
Interaction with Universities/ RD Institutions/
Academics

• Universities
• Harvard University
• Stanford University
• University of Delhi
• University of New Castle
• University of Texas
• Social Science Research Centre, Berlin
• University of East Anglia
• Maryland University
• University of California (Santa Cruz)
• Development Institutions
• The World Bank
• USAID
• DFID
• UNDP
• Indian Research Institutions
• Rajiv Gandhi Institute of Contemporary Studies
• 3I Network of IIM (Ahmedabad), IIT (Kanpur) and
  IDFC

5
Raftaar.comFirst integrated search engine in
Hindi
6
Raftaar.comFirst integrated search engine in
Hindi

• A simple user interface to type in Hindi
• (A primary school dropout can also use)
• Font Hassle Free Gets information from sites
  ir-respective of the font
• Regular update of index for latest and most
  relevant information

7
Indicus IT

• Software
• Software for gauging competitiveness of SSIs -
  UNDP/FICCI
• Consumer Tracker - Purchased by Maruti, Bata,
  Parle ITC etc.
• Market Skyline Purchased by more than 300
  corporate
• Indicat socio-economic data analyzer and
  mapping software
• User Needs
• Largest Online survey of net users in India
• Quarterly survey of consumer preferences
• All India offline survey of internet users
• Continuous user monitoring through raftaar

8
Our Roles

• User Surveys to understand needs
• Continuous Inputs for development of commercially
  viable products even at intermediate stages
• Developing large and varied samples for testing
  engines
• User Feedback Analysis to ensure better
  development

9
Would Like to Develop and Deploy

• An online tool to provide basic transliteration
  of web sites with English content.
• An online tool to provide basic translation of
  web sites with English content.
• An online tool to provide basic translation of
  terms in data web sites such as RBI, Agricoop,
  Indiastat, etc.
• In the process ensure development and
  availability of Hindi content

10
Business Volume/ Financial Details
Indicus Analytics

• Indicus Netlabs Pvt. Ltd. has been spun off
  recently from Indicus Analytics
• Indicus Analytics has been in existence for 5
  years
• Five fold increase in revenues in as many years
• Revenues for 2004-05 gt Rs 1 Cr.
• Consistently increasing profitability (net
  profit/revenues)

11
Technical Capabilities

• Over 15 years individual experience in Software
  Industry
• Over 5 years experience in research oriented work
• Over 5 years experience in database analysis,
  management and development
• First Integrated Hindi Search Engine- www.
  Raftaar.com
• Comprehensive understanding of language related
  complexities
• Experience across various technical platforms,
  languages and environments

12
Machine TranslationMethodology
English document
Input
Machine Translation
Transliteration
Corpus
POS Tagging Engine
Bilingual dictionary
Output
Chunking Engine
Morphological Analyzers Generators
Hindi Document
Transfer Engine
RulesDatabase
Evaluation of MT System
13
Machine TranslationMethodology
14
MT Development Capabilities

• Understanding of Fonts and the associated glyphs
• Experience in mapping of all the alphabets for
  Devanaagari (Hindi) for majority of the currently
  used fonts
• Development of tool to understand PDF based
  information
• Experience in working with Unicode and INSROT
• Experience in developing a corpus in Hindi which
  is in Unicode. Our current web crawlers are
  continuously updating the corpus.

15
Decision Process

• Information Collation
• Brain Storming with Team
• Decision Alternatives
• Decision taken by assigned entity

• Technical P. Srinivasan
• Operational P. Srinivasan, Peeyush Bajpai
• Organizational Peeyush Bajpai Laveesh Bhandari
• Financial Laveesh Bhandari

16
Manpower Involved

• Project Advisor Dr. Laveesh Bhandari
• Overall Operations Peeyush Bajpai
• Technical Development P. Srinivasan
• Development Team A combination of software
  engineers researchers
• Coordination Mamtesh Kumar
• Business Development Peeyush Bajpai
• Media Networking Kapila Chaplot

17
Marketability
18
Marketability

• Product/ Service
• Demand
• Value Proposition
• Client Orientation
• 5 Ps of marketing
• Credibility
• GOI
• IIT
• CDAC
• Indicus and its networks with policymakers,
  media, and corporate

• Many of our studies have been released and
  referred to by eminent Indians including the
  President Dr. APJ Abdul Kalam, the Vice-President
  Shri Bhairon Singh Shekhawat, the Prime Minister
  Dr. Manmohan Singh, the Finance Minister Dr. P.
  Chidambaram, the Panchayati Raj Minister Shri
  Mani Shankar Aiyer, former Deputy Prime Minister
  Shri L.K.Advani, and many others

19
Marketability

• Media Coverage
• Hindi Dailies Dainik Jagran, Hindustan,
  Jansatta, Prabhat Khabar
• English Dailies Indian Express, Telegraph,
  Chronicle, Times of India
• Business Dailies Economic Times, Business
  Standard, Financial Express
• Magazines India Today, Outlook
• Past Clients (one example each)
• The Government The 12th Finance Commission
• The Media India Today
• International Academia Stanford University
• Development Institutions The World Bank
• International Aid Organizations DFID, Government
  of UK
• NGOs Liberty Foundation
• Indian Research Institutions Rajiv Gandhi
  Institute of Contemporary Studies
• Networks 3I Network (IIM A, IIT (Kanpur) and
  IDFC)
• Companies Hindustan Lever
• Associations Confederation of Indian Industry