Google Summer of Code 2016
Indic Project participated in the Google Summer of Code program this year and we had 7 students working with us for months as a part of it.
This was the first time Indic Project participated in GSoC. We applied as SMC till last year and had been a participating organization for 4 times, last in 2014.
There were 7 projects in total, of which 4 were enhancements to existing projects.
Sreenadh worked on speech recognition for Malayalam under the guidance of Prof Deepa. For this he built a language model, an acoustic model, and a phonetic dictionary for Malayalam. These can be used with CMU Sphinx to help in accurate speech recognition. The acoustic model helps in transcribing sound into phonemes. The phonetic dictionary can then be used to convert those into words. The statistical language model gives the probability of which words appear together and thus helps in extracting meaningful sentences. His repository saw collaborative effort by a lot of people, including his little brother. Read Sreenadh's final report.
Jerin has built a sandhi splitter for Malayalam based on machine learning. His also worked on improving the spellchecker module of libindic and adding a REST API for libindic. Vasudev and Hrishi were his mentors. Read Jerin's final report.
Balasankar who's a long time volunteer of SMC, worked an inflection aware spellchecker too using an approach based on stemmers. While on it, he reorganized the modules and improved the libindic project structure too. Hrishi was his mentor. Read Balasankar's final report.
Vishnu worked on adding Varnam support into Indic Keyboard. Although the JNI wrapper based approach didn't work as planned, he completed his project by making use of the REST API. Navaneeth and Jishnu were his mentors. Read Vishnu's final report.
Arushi worked on better user on-boarding for the Indic keyboard, giving the user hints about what locales might not be supported by their device, helping them choose a theme, etc. She also worked on adding Santali and Soni layout. Jishnu mentored her. Read Arushi's final report.
Irshad worked on machine transliteration, allowing for cross transliteration among all Indian languages. His brother Riyaz and Santosh were his mentors. Read Irshad's final report.
Anwar worked on enhancing iBus-Braille. He built a web-based braille input tool and made a lot of useful enhancements to the project. Nalin and Samuel were his mentors. Read Anwar's final report.
Most of the GSoC participants are working with our community after the GSoC period too in solving bugs and adding enhancements that they hadn't found time for during the GSoC period. If you are looking forward to contributing to Indic Project, get in touch with us by joining our IRC channel #smc-project on freenode.