Rafiya Begum Email & Phone Number

Hyderabad, Telangana, India

(1)Worked on Data Migration on AWS S3

Hyderabad, Telangana, India

Hyderabad Area, India

Bangalore

• Text Normalization: Worked on Text Normalization for TTS (Text to Speech) for Telugu language, where things such as cardinal numbers, fractions, decimals, ordinal numbers, dates, time, punctuations, etc., which are written in non-text form are converted to text form first and then to speech. In a specific language, each written form of an alphabet of a language is mapped to a sound and since numbers don’t have a mapping to sounds, the number data is converted to text form and then this text gets mapped to sounds. The number data is very exhaustive hence, this conversion from non-text form to text-form is an extensive work. This conversion is done on a Text Normalization tool. The tool also gives us the option to test these conversions so I prepared some test cases for each topic (cardinal numbers, fractions, decimals, ordinal numbers, dates, time, punctuations, etc.) of conversion and ran these test cases in the tool to find out the errors. Then the error cases were fixed using the tool. After this, the testing was also done at the front-end and similarly, if there were any errors, they were fixed at the back-end.• TTS Front-End: Worked on the mapping of the Telugu graphemes (written form of alphabets) with the Telugu phonemes (units of sound). Worked on the Syllabification (breaking of words into segments) of Telugu words.• Also trained other language experts on Text Normalization and Automatic speech recognition work.

Bangalore

I have worked on Text Normalization for TTS (Text to Speech) for Hindi, where things such as cardinal numbers, dates, time, punctuations which are written in non-text form are converted to text form first and then to speech. The other task that I worked on is Inverse Text normalization for ASR (Automatic Speech Recognition) for Hindi where I converted the cardinal numbers, dates, time, punctuations from spoken form to written number form. Tested the above data to fix the bugs and ambiguities. Did a Front-end testing to check the overall performance of the speech. Reviewed the Hindi Script for Text-To-Speech Recordings. This work involved reporting of pronunciation and grammar errors at the Language side and also to see if the recordings are according to the script. Also listened to the recordings and judged the recordings based on certain parameters like Voice volume level, Pitch and Pace, Loudness, Noise Volume Level, etc., at the Non-Language side using Audacity tool.

Bengaluru Area, India

I have analyzed Code Switching from the linguistics point of view. Code Switching (CS) means switching between two or more languages while communicating. For computational processing of code-switched data, we need annotated data for understanding, processing, and generation of code-switched language. For this purpose, we came up with an annotation scheme for annotating the pragmatic functions of CS in Hindi-English (Hi-En) code-switched tweets based on a linguistic analysis and some initial experiments.

Hyderabad Area, India

--Have been associated with the Hindi TreeBank Project. --Visited University of Colorado (Boulder, Colorado state, USA) for a period of less than one month on their invitation to work on the Hindi TreeBank Project.

Ph.D., Computational Linguistics

Activities and Societies: Guiding Masters students, research in language technologies•Created Indian language resources which are used.

Master'S Degree, Linguistics

Activities and Societies: Korean Language Certificate CourseM.A. (Linguistics)

Ph.D, Computational Linguistics