| |
Dr. Farah Adeeba is Manager Research at CLE, working in the area of statistical language processing for past 10 years. She is Recently she has complated her doctorate in computer science. Her Ph.d is on spoken language identification using acoustic features and machine learning approaches. She has developed domain/genre identification system, and developed a highly efficient word formation system using statistical techniques for the Urdu OCR currently being finalized for release by CLE. During this project she was responsible for management of system efficiency and post processing by employing statistical language processing techniques. Currently, she is working on the text classification project and responsible for management of system performance and improved accuracy by employing feature selection techniques. In her MS Thesis she worked on word sense disambiguation system to improve automatic machine translation accuracy. In the past she has also worked on video summarization by using shot-boundary detection, key-frames selection, and face detection. She has also worked on the creation of linguistic resources and participated in the design, collection and cleaning of text corpora of different sizes i.e. 100K, 500K, and 1M in a project funded by DAAD Germany. In corpora development task she was also responsible for developing guidelines and utilities for completion of the task in a systematic way. She has also managed the syntactic annotation of these text corpora. Along with the creation of linguistic resources she has developed language processing utilities i.e. corpus annotation, automatic part of speech tagging, corpus cleaning, and sense tagging. She has also participated in the development of WordNet. |
|
