MSR India
Microsoft Research's newest lab in Bangalore, India - An article interviewing Kentaro Toyama, assistant director of Microsoft Research in Bangalore, India. I like the vision that they have for research in India. An excerpt -
The six research areas we're currently focusing on are photography; digital geographies, which includes any kind of digital map or location-based services and software; multilingual systems, including speech recognition, natural language processing, and building systems that interact across different languages; hardware for communications, including distributed sensor networks; software engineering, which looks at creating tools that make software development easier; and emerging markets, or how computing will impact social and economic development.
All these ideas are interesting and worth exploring and some of them (like multilingual systems) have a huge potential in a place as diverse as India. I also feel that the socio-economic disparity will slow the pace for most of what they envision. Though this disparity might soon be disappearing - you never know!.
Anyway, I particularly am interested and like the idea of being able to search and return results without having any kind of language barrier that exists today. Makes me wonder - Will the problem of imprecise user queries decrease if the search engine was adapted to the user's language? Will we need novel interfaces to support these different languages? Will the searching, indexing etc. be done in the target language itself or assuming that we have cross-language ontologies, will we convert/translate from that other language to English, search in English and then translate back to that language? (This is something similar to what happens in Speech Recognition (SR) today - speech is converted to text, you search on this indexed English text and then return the corresponding audio/video files containing that text). What kind of information will we lose in switching back and forth between languages? Will the kind of information lost, be different if we switched from say Marathi to English as compared to French to English or say Russian to Chinese?
The six research areas we're currently focusing on are photography; digital geographies, which includes any kind of digital map or location-based services and software; multilingual systems, including speech recognition, natural language processing, and building systems that interact across different languages; hardware for communications, including distributed sensor networks; software engineering, which looks at creating tools that make software development easier; and emerging markets, or how computing will impact social and economic development.
All these ideas are interesting and worth exploring and some of them (like multilingual systems) have a huge potential in a place as diverse as India. I also feel that the socio-economic disparity will slow the pace for most of what they envision. Though this disparity might soon be disappearing - you never know!.
Anyway, I particularly am interested and like the idea of being able to search and return results without having any kind of language barrier that exists today. Makes me wonder - Will the problem of imprecise user queries decrease if the search engine was adapted to the user's language? Will we need novel interfaces to support these different languages? Will the searching, indexing etc. be done in the target language itself or assuming that we have cross-language ontologies, will we convert/translate from that other language to English, search in English and then translate back to that language? (This is something similar to what happens in Speech Recognition (SR) today - speech is converted to text, you search on this indexed English text and then return the corresponding audio/video files containing that text). What kind of information will we lose in switching back and forth between languages? Will the kind of information lost, be different if we switched from say Marathi to English as compared to French to English or say Russian to Chinese?
posted by anonymous at
2:58 AM
1 Comments:
Microsoft is doing a wise job in doing early studies about the Indian market and tapping it quickly, thereby getting an early lead and a head-start...
Cheers
Varahasimhan.
By
Anonymous, at 9:43 PM
Post a Comment
<< Home