SolutionsSolution Frameworks from TCS Innovation Labs - Mumbai Speech Script Image and Natural Language
Wireless Systems and ApplicationsSpeech Script Image and Natural LanguageThe speaking style and articulatory capability of spoken language of a person is assessed very easily even by a person who is not familiar with the language of the spoken speech. Work is directed towards building a platform having similar capability. The current system can automatically judge the accent of a person by listening to a set of predefined words and sentences. The speech samples (of all the words and sentences) of the person are compared with the corresponding statistical models (continuous HMMs) derived from the training speech data. Experiments performed on 30 candidates show that the system is able to match the judgment of the accent expert in more than 90% of the cases. In another experiment the performance of the system was analyzed against two different human experts who independently carried out the classification process. Results showed that the HMM based accent analysis system agrees better with the two experts than the agreement between two human experts. (Reference: An Approach towards Automatic Evaluation of Accent and Style PVS Rao, Sunil Kopparapu, COCSDA, New Delhi, Nov 2004.) A speaker verification system based on MFCC and LPC parameters has been in use in ATAG and several other offices for several years. The system enables verification of the identity of a person based on their voice signature. Use of this biometric-based technique for verification on the basis of an individual’s voice covers a wide spectrum of applications. The system has been tested for speakers speaking over noisy telephone channels. Speakers can choose any password of their choice and need to repeat their chosen password a few times over a period of time for enrolling with the system. The system learns and captures the voice signature by building statistical models from the training data set of the user. Online Answering System with Intelligent Sentience is a question-answering answering system enabling man-machine interaction platform in natural English. It is specifically oriented towards answering questions and providing information on a given topic. The system has been designed to work as a human would do in an oral transactional interaction. Like a human, the system does not parse the sentence but rather understand the intent of the question by paying attention to only key concepts. The QA system does not parse sentences to understand the intent of the question. It is based on extraction of key concepts and key words. This flexibility gives OASIS a handle to work with multilingual inputs with minimum architectural changes. This basic platform has been used in several QA implementations. SMS based Yellow Pages Information Server Yellow Pages are directories that source information about various commercial organizations like their addresses, phone contact and other details. Until recently, the only way to access these yellow pages directory information was to physically look into a huge hard-copy directory, which was not only laborious but also time consuming and required the user to be familiar with the organization of the directory. More recently, there have been IVR based contact centers that have been set up which can be used by the users to query information. While it is easier than browsing through the physical directory, it still has several pitfalls. The time spent on trying to get the information is quite large and at the end of enquiry one is not sure if one will get the information that one is looking for. A novel system was built using the base QA platform (which has been implemented for a major telecom operator) to access yellow pages directory information on the mobile phone by sending a short message service (SMS). The central idea of the proposed method is to avoid any constraint on the way the user can query the yellow pages directory except that it be in natural English. The system, which uses natural language processing (NLP) techniques, understands the intent of the query and intelligently searches the yellow pages directory to retrieve information. In case exact answer is not found, the system retrieves information which is closely associated (in the human sense) with the query. This retrieved information is then sent back to the user in the form of a SMS. (Reference: Accessing Yellow Pages Directory Intelligently on a Mobile Phone Using SMS by SK Kopparapu, AC Srivastava, S Das, R Sinha, M Orkey, V Gupta, J Maheswary and PVS Rao MobiComNet 2004 Vellore.) A system to extract information automatically from resumes and convert them into a structured database has been developed. The system is based on a set of natural language techniques to derive the required information for the resume. A user interface to extracts resumes of interest has been developed. Work is on to enable a natural language interface to extract resumes satisfying a specific requirement from the database. Fetching information form even a well designed website even for a well informed user is tedious process. To obtain the desired information, one is required to select from several drop-down menus, or click on several hyperlink and manual browse the displayed information. We have developed a natural language interface (NLI) to the Indian railway website which enables the user easy and succinct information retrieval by posing queries in natural English. The system has been piloted internally. An interface has also been enabled to give the user the facility to get information from the Indian Railway website through a mobile phone by sending an SMS. (Reference: A Natural Language Interface for a Railway Website Sunil Kopparapu, Akhilesh Srivastava, PVS Rao Second National Conference on Innovations in Information and Communication Technology 2006, 7-8 July, PSG College of Technology - Coimbatore.) KisanMitra – QA for Rural India Farmers in most rural areas in India not only need expert and timely suggestion to obtain rich harvest of their crops but also need information regarding the subsidies, government schemes to make cultivation pay rich dividends. Expert guidance comes in the form of a human expert visiting the village and the farmers being able to get their turn to seek answers to their queries. The developed system can act as an expert and answer queries of the farmers. This QA system has been christened KisanMitra, friend of the farmer. The idea in building this system is to give access to information 24 x 7, to keep the information that reaches the farmer updated, enable the farmer to query in his own language without being strict on grammar or construct of the query. The system is intelligent in the sense, it understand the intent of the query and provides responses. In the absence of exact answers not being present in its KisanMitra, it provides answers which are close in some (human) sense. This is being developed as part of TCS corporate social responsibility initiative. (Reference: KisanMitra: A Question Answering System for Rural Indian Farmers Sunil Kopparapu, Akhilesh Srivastava, PVS Rao International Conference on Emerging Applications of IT (EAIT 2006) Science City Kolkata, February 10-11, 2006.) Wireless Systems and ApplicationsFarmers in India, especially marginal farmers, need information and specific advice specific to their activities. Proper dissemination of farmer-related information and advice would lead to effective usage of fertilizers and pesticides, which in turn would increase productivity and efficiency, reduce environmental damage, and improve the economic status of farmers. Farmers usually ask the following questions in different parts of India: How is the quality of my soil ? Which crop should I grow and which fertiliser should I use? Will it rain in the next few days? Should I sow seeds now? mKRISHI tries to address these concerns of the farmer with the help of CDMA technology. A consortium of eco-partners from various fields is working with TCS for mKRISHI deployment. Using CDMA technology, handsets and other innovations like mobile cameras, soil sensors, and automatic weather stations make it possible to send farm-related information to the expert over cellular network. The expert's advice can then be sent to a farmer's handset in a local language. This expert advice combined with local weather prediction can be a great boon for farmers. This integrated system will assist farmers achieve a better standard of living and will also eventually lead to a knowledge-based economy. The Weather application is a micro climate prediction application and gives a seven day prediction of Temperature, Precipitation and Cloud Cover in the village from which the query is sent. All this information is conveyed to the farmer in his native [local] language and for his region. The Pesticide advisory application is the second application in this suite. It allows the farmer to capture an image of the crops and request advice from central pool of experts. The advice is then conveyed to the farmer in native [local] language. Mandi price information, provides crop price information to farmers from nearby mandis, in native [local] language. Solutions on the PIM2R FrameworkWe have developed various solutions based on our Patented Technology Packet Interactive Multimedia Response (PIM2R) like Unobtrusive Advertisement Push (UAP) The mobile handset screens can be a breakthrough medium for displaying advertisement images and information. No media footprint can match the number of users the ads can be targeted to, using mobile phones. However, most mobile users will not like obtrusive push of the information/images on their mobile screens. Purposes solved:
Mandi Ka Bhaav– This application gives information on the latest Spot and Future Prices of various commodities and reads them out to the user in a language selected by him. The information is fetched from the NCDEX server. View Demo Mixed Bag Contests– This application is a Contesting portal which is capable of playing multimedia questions based on pictures, videos, audio, text to the contestant and accepting user inputs in numerical form. We are developing an instance of this application for TTSL along with Mobile2Win called VoiceBox. Flash Cards on Mobile– This application is useful to any student who is giving entrance tests like GRE, TOEFL, CAT, etc. It presents over 3500 words in wordlists to the user with the pronunciation of words. Location Based Solutions: We have developed two solutions for TTSL using CDMA Sector Information. These solutions can be extended to accommodate Assisted GPS Solution once it is deployed in India. Friend Locator – This application allows you to track your friends current location and has features like Maintaining a Friends list, Adding & Deleting friends, Chatting with Friends through SMSes, Updating your own location, Hiding / Unhiding yourself, etc. Bus Tracker – This application essentially allows a parent to track a child’s location while he travels in his school bus. The application has two clients – one with the driver which constantly updates the location of the bus and the other with the parent which receives alerts on the bus location. |