|
[ << Go Back ] |
[ ^^ Goto Home ] |
|||||||||
|
Rohit Kumar
Research Engineer, Speech Technology
Group,
[ Education ] [
Research Experience ] [ Research Projects ]
[ Publications ] Career Objective To pursue and collaborate in research, development and teaching of technologies at academic / research oriented institutes and organizations and to work towards innovation of products for mass use Educational Objective To join and pursue a Ph. D. program specializing in Speech Technology and Spoken Language Systems in an internationally reckoned group with significant contributions to this field Personal Particulars
Male, Born on 09 January 1982
Postal Address: LTRC, IIIT Hyderabad, Gachibowli, Hyderabad, INDIA 500 019
Email:
rohit [AT] iiit [DOT] net, rohit [DOT] kumar [AT] gmail [DOT] com [ Top ]
Bachelor of Engineering (in Computer
Science): August 1999 – May 2003
Intermediate (Junior College): June 1997
– May 1999
Matriculation: May 1996 – April 1997 Some of the several courses takeN
[ Top ] Tests TAKEN:
July 2003 - Presently: Research Scientist in Speech Technology Group of Language Technologies Research Center (LTRC), International Institute of Information Technology (IIIT), Hyderabad, INDIA. Involved in development and improvement of Text – to – Speech Systems and related Applications for Indian Languages. Also coordinating activities related to partnership between PICOPETA and IIIT Hyderabad for developing system for Simputers. December 2002 – January 2003: Winter Internship at LTRC, IIIT Hyderabad. Contributed to development of Unrestricted Domain Text to Speech Systems for Indian Languages and Low Memory Device Synthesizer June 2002 – July 2002: Summer Internship at LTRC, IIIT Hyderabad in Summer 2002. Contributed to development of Unrestricted Domain and Limited Domain Text – to – Speech systems and developed speech output applications for Simputer Areas of Interest · Speech Technologies · Spoken Language Systems · Natural Man/Machine Interfaces · Natural Language Processing · Artificial Intelligence · Computer Graphics [ Top ]
Languages:
C / C++, Python, QuickBASIC, LISP, PERL, HTML,
Operating Systems: Linux, Windows NT/9X, DOS (core), SCO Unix, Solaris Hardware Platforms: PCs, Sun E - 450 Servers, Simputer (Strong ARM) Toolkits/ Libraries: Qt, GTK+, STL, Sockets, Festvox Framework · Sound knowledge of Data Structure, Algorithms, OOP Design Patterns · Practical Exposure of administering networks at school, college and hostel · Strong Debugging Skills [ Top ] Automatic Speech Recognition Systems for Indian Languages
· Current activity is to setup resources and infrastructure to collect Speech Corpus for targeted languages · Collected Text Corpuses for targeted languages and made them available in required notation and formats for Optimal Text Selection. · Worked with phonetics of targeted languages. · Developed a telephonic Interactive Voice Response application using Dialogic for collected speech corpuses over telephone line. Khabrein: Online Hindi News Channel
· Developed an Indian Language News Search Engine · Several issues like Font and Format Unification, Document Classification, Web site specific crawlers have been addressed. · Web based Speech Out Interface GASpSynthesizer
· Developed and Experimenting with a Genetic Algorithm based approach for Unit Selection in the framework of Concatenative Speech Synthesis · Choice of Appropriate Fitness Functions; Size of Initial Population; Implementation of Selection, Combination, Elitism and Mutation operators are some of several issues that need to be experimented with Unit Pruning in Speech Database without loss of Naturalness
· Developed an approach to prune away units which do not contribute to prosodic coverage of the units or the prosodic coverage provided by which is rarely required (statistically) · Experiments to find the optimum extent of pruning without significant loss of naturalness are underway Rule – based Approach for Building Non – Native Pronunciation Lexicon Independent Research Location: Punjab Engineering College, Chandigarh Duration: February – May 2003 · Correspondence between Native and Non Native Pronunciations of English observed and rule format for modeling the differences has been proposed · Generic Algorithm for mapping Graphemes to Phonemes in a word’s pronunciation developed. · Experimented with Example based Approach for automatically building rules PECMail: Email Client for Hindi (with Speech Support)
· A basic Email Client with Support for sending and receiving mails in Hindi besides English · Common features of Email clients like POP3 access and SMTP implementation, compose, reply, forward, etc. where implemented · Incoming mails in Hindi are read out by the client · A GUI based keyboard for typing Hindi for users unaware of Hindi Keyboard Layout Low Memory Device Synthesizer
· Contributed to development of a Low Memory Device Synthesizer (LMDS) for Simputer enabling speech synthesis for Indian Languages (currently Hindi and Telugu) · Automated methods for selecting the best units from a given speech database were conceived and experimented with · Good Quality Speech Synthesis is achieved for Database Size as small as 1.4 MB · Recently, API for using this system have been developed. Samachaar Vaani: News Reader Project Guide: Mr. Sanjeev Sofat Location: Punjab Engineering College, Chandigarh Duration: 7th Semester, August – December 2002 · Developed a News Reader Software that reads out latest news from BBC Hindi News Websites · A flexible Server / Reader Framework was proposed that would enable any news service provider to provide spoken news service using Automatic Speech Synthesis hence avoiding lot of manual costs involved in providing spoken news service Indian Language Speech Synthesizer based on Data Driven Approach [ See Demo ]
· Contributed to development of a Generic Unrestricted Domain Synthesizer for Indian Languages using Syllable level Unit Selection & Concatenation on lines of the Limited Domain Synthesizer · Used a Prosodic Mismatch Function to select the most natural (prosodically harmonious) sequence of units in given phonetic contexts · Several Text Processing Issues related to Indian Languages like Syllablification and Inherent Vowel Suppression were discussed and implemented Continued System Development (August 2003 onwards) · More recently, a robust API for the Indian Language TTS has been implemented to facilitate application development using the system in Windows · Developed of font converters and text processing modules among various Hindi Fonts, Unicode, ISCII and other notations Limited Domain Speech Synthesizer Synthesizer
· Contributed to development of a Generic Limited Domain Synthesizer for Indian Languages based on Syllable level Unit Selection · Related Applications o Talking Tourist Aid Adapted a Talking Tourist Aid Application for Hindi and Telugu to use the Limited Domain speech synthesizers and ported it to Simputer o Talking Clock Adapted a Talking Clock Application for Hindi and Telugu to use the Limited Domain speech synthesizers and ported it to Simputer Deepti: Computing Time Companion Project Guide: Mr. Sanjeev Sofat Location: Punjab Engineering College, Chandigarh Duration: 6th Semester, February – May 2002 · Developed an Intelligent Hindi Speaking bot with speech out capability using diphone databases · Used AIML Engine for providing Natural Language Dialog capabilities to the bot · Please see: http://deepti.nourl.org/ · The Project was featured on BBC Technology Review Show “Go Digital” and in several other national publications Speaker Verification: Speech based biometric Authentication · Implemented a speaker verification algorithm based on rate of positive zero crossing for use in project on telephonic authentication of speaker as the software component of major project of one of my seniors Several other projects were undertaken as a part of practical work of various courses. For a exhaustive list of these please see http://www.geocities.com/rohitofpec/projects.htm [ Top ] Journals:
Conferences:
"Automatic Pruning of Unit Selection
Speech Databases for Synthesis without loss of Naturalness"
Accepted at International Conference on Spoken Language
Processing (ICSLP) 2004,
"A Genetic Algorithm for Unit
Selection based Speech Synthesis"
Accepted at International Conference on Spoken Language
Processing (ICSLP) 2004,
"Unit Selection Voice for Amharic
using Festvox" Proceedings of 5th ISCA Speech Synthesis Workshop (SSW5), Pittsburg, June 2004
"Building Non - Native Pronunciation
Lexicon for English using a Rule based Approach"
Presented at International Conference on Natural Language
Processing (ICON) 2003 Earlier Accepted at:
“Implementing a Natural Language
Conversational Interface for Indian Language Computing”
“Samachaar Vaani: A Framework for
providing Automated Spoken News Service”
Sixth International Conference on
Information Technology (CIT), 2003
“Experiments with Unit Selection Speech
Databases for Indian Languages”
“A Data-Driven Synthesis Approach for
Indian Languages using Syllable as Basic Unit” Technical Contributions to Student Publications and Contests in College:
“A Framework for E – Governance in India
with Local Language Support”
“Observing and being a part of Evolution
of Language: Some Random Thoughts”
“Data Driven Approaches: A Basis for
Everything” [ Top ] Professional Memberships and activities
I E E E Computer Society
I E E E · Member of The Indian Science Congress Association ( I S C A ) for the year 2003 [ Top ] IEEE Computer Society Richard E Merwin Student Scholarship 2003 [ Award Website ] · Every year upto four students world over are selected to be recognized and rewarded for active leadership in IEEE Student branch chapters · Please see http://www.computer.org/students/schlrshp.htm#merwin for details about the scholarship and selection criteria
· The silver medal for Best Final Semester Project in Department of Computer Science and Engineering awarded to me during Annual Convocation
· 1st prize in Pre-defined software contest in PECFest 2001 (PEC technical festival 2001) · 1st prize in On-the-spot software contest in PECFest 2001 · 2nd prize in Debugging contest in PECFest 2001 · 1st prize in Pre-defined software contest in Panoplia (PEC technical festival 2000) State Science Exhibitions · 3rd prize in State Science Exhibition organized by State Institute of Education, UT, Chandigarh, in 1998 for projects on Traffic control by computers · 1st prize in State Science Exhibition organized by State Institute of Education, UT, Chandigarh, in 1997 for projects on Role of IT in Education Technology [ Top ]
· Co – opted member of College’s Student Council 2002 – 03 for excellence in technical activities · The Principal nominates upto four co – opted members for excellence in various circles
· My responsibilities included the aspect of uploading, maintenance and administration of the website
· PECMag, Souvenir and VISTA cover pages for the year 2002 – 03 were designed by me
· SemantiCa, an Intra College On-the-Spot programming contest is now the premier event organized annually by IEEE Computer Society Student Branch Chapter at PEC. It was conceived and implemented for the first time by me
· Due to my contribution to the activity over years, I was appointed co – convenor of the event during 2001. It provided me with first hand experience in event management and organizing human and material resources
OTHER INTERESTS Cycling, Music, Movies, Badminton, Hiking REFEREnces
Prof. Rajeev Sangal
Mr. Sanjeev Sofat
Mr. S. P. Kishore
| ||||||||||
|
[
Top ]
This Page Is:
http://speech.iiit.net/~rohit/cv.htm | ||||||||||