Scientists at IIT Kharagpur are developing resources and tools to help computers provide a better online experience to an user…
IIT Kharagpur (Photo: Facebook)
Scientists at IIT Kharagpur are developing resources and tools to help computers provide a better online experience to an user to read Bengali texts and find answers to his queries in the same language.
Prof Sudeshna Sarkar, head of the department of Computer Science and Engineering at IIT-KGP said, currently the institute is working with Google to develop a Bengali treebank which is an important resource to understand the grammatical structure of Bengali texts as well as their meaning.
Sarkar gave examples about how their final product would work.
Advertisement
“If you wish to know about flood related reports in West Bengal that have appeared in various Bengali newspapers over the past 10 years, the computer can prepare a report for you with dates and locations of the events as well as the steps taken by the government,” she said.
Likewise if an end user wants to know what several Bengali authors since the time of Rabindranath Tagore had said about Kolkata in their works, the computer can give the answer, she said.
“We are trying to create tools and resources for understanding Bengali by the computer. Such tools will enable a far better online experience for a Bengali language user and give him better access to the information available in Bengali texts,” Prof Sarkar said.
The researchers at the Computer Science and Engineering department are developing resources and tools for Bengali, which will facilitate word level as well as sentence level grammatical analysis of Bengali texts.
“The research will enable better understanding of Bengali language so that the computer can retrieve answers to your questions,” she said.
“We had been involved in a project called Sandhan partnering with several institutions to develop a cross-language Indian search engine. Typing a query in Bengali, the system will also let you access relevant documents in other languages such as Hindi and English,” she said.
The 'Sandhan' system also handles Marathi, Gujarati, Tamil, Telugu, Odia and Assamese.
Advertisement
