Fuad. Rahman has extensive background in natural language processing, computational linguistics, pattern recognition and text Analytics. He is a technologist with ten years of experience in building research based software products, with a demonstrated ability to drive product planning, development and launch. He has worn many caps, from coding, to architecting, to leading R&D projects and successfully migrating research to commercial products. He has direct experience leading the charge with customer interaction - flagging customer issues, serving as a primary contributor to cross-functional release teams, keeping up with industry/competitive trends and fostering relationships with key partners.
Abstract: In this paper we describe a preliminary, work-in-progress Spoken Language Understanding Software (SLUS) with tailored feedback options, which uses interactive spoken language interface to teach Iraqi Arabic and culture to second language learners. The SLUS analyzes input speech by the second language learner and grades for correct pronunciation in terms of supra-segmental and rudimentary segmental errors such as missing consonants. We evaluated this software on training data with the help of two native speakers, and found that the software recorded an accuracy of around 70% in law and order domain. For future work, we plan to develop similar systems for multiple languages.
Abstract: A novel multiple-expert framework for recognition of handwritten characters is presented. The proposed framework is composed of multiple classifiers (experts) put together in such a manner as to enhance the recognition capability of the combined network compared to the best performing individual expert participating in the framework. Each of these experts has been derived from a novel neural structure in which the weight values are derived from Clifford algebra. A Clifford algebra is a mathematical paradigm capable of capturing the interdimensional dependencies found in multidimensional data. It offers a technique for concise data storage and processing by representing dependencies between the component dimensions of the data which is otherwise difficult to encode and hence is often employed in analyzing multidimensional data. Results achieved by the proposed multiple-expert framework demonstrates significant improvement over alternative techniques
Abstract: A linguistic tutor to help SMEs to resolve ambiguities, understand the semantics of slang and double meaning constructions for a foreign language text.
Abstract: BCL Knowledge Management Center (KMC) is a knowledge based document and information management platform. This offers universal information access for the Mobile Warrior.
Abstract: To make any electronic document of any format universally accessible to the Mobile Warrior from any electronic device, including handheld Personal Digital Assistants (PDAs) using wireless connections
Abstract: The traditional approach to the specification of practical high-performance classification systems has been to focus on the implementation of powerful individual algorithms to address the problem. There has been an increasing recognition, however, that individual stand-alone classifiers are often not sufficiently robust to deal with the huge degree of variability present in many types of data, and a multi-classifier approach is now commonly adopted to deal with particularly difficult classification problems. This approach can make use of the principle of complementarity and can exploit the strengths of certain recognition algorithms while avoiding the weaknesses of others in relation to a particular data domain.
In this project we aim to develop and evaluate novel structures for the effective analysis and subsequent synthesis of systems involving the fusion of multiple classifiers. To illustrate the generic nature of this approach, we use a variety of data from different domains reflecting the diversity of characteristics met in real world applications. The work is being carried out using handwritten characters and biometric data to define typical applications, as well as other publicly available databases where appropriate.
Abstract: Dealt with ways of combining multiple 'classifiers' or 'software experts' to deliver a robust, consistent and highly reliable classification performance.
