Tamil Computing Research Experiences- Part 3
In my earlier posts of this series, I had written about my Tamil computing experiences I had during my JRF job at Anna University and during my PhD. This post is on my post-PhD Tamil computing research experiences that I had & having at SRM. Today would be so apt to write on this, as I complete 5 years at SRM by today. It has been a joyful learning-teaching-researching journey so far and I really wish to write a post on this in my non-technical blog soon but I wanted to finish this series today so landed here.
It is a sheer luck when your workplace appears to be the one you had imagined. Yes, SRM did fulfill my teaching and research dreams. I still remember my first day afresh and I was asked to choose subjects to teach. All subjects were already opted by the rest of the faculty members and I was left with very few choices . I wanted to opt Natural Language Processing (NLP) but in the few choices given to me did not include NLP. I asked if NLP is available. That faculty looked into the list again and said yes NLP is there . This was the first positivity I absorbed at SRM.
To my dismay, there were only 20 students who had opted NLP and no one was a Tamil student as at SRM, we have students almost from all states of our country. My first batch of NLP students are awesome till date :) That's when I thought I had to start my independent research on Tamil computing and teaching NLP did help! Thanks to my then HOD, Dr.Poovammal Mam , current HOD, Dr.B.Amutha and all Professors who encouraged and believed a lot in me. Poovammal mam initiated a new research lab called, "Natural Language Processing and Machine Learning lab" soon after I joined and shifted my room to this lab. Amutha mam made me the head of the Special Interest group called,"Natural Language Processing and Regional Languages". This was all due to the fact that NLP was a rare domain then and it is no more now :) Dr. Revathy Venktraman mam, Dr.Pushpalatha mam, Dr. C.Malathy mam and Dr.D Malathy mam have also assigned me duties which helped to grow as a good academician as I lacked academic experience before SRM.
I had to wait for 2 years post my PhD graduation to get my PhD supervisor recognition. I was given a project on "semantic search engine for Indian languages" done by my NLP students funded by SRM but Tamil was still missing so thought to extend my PhD research work on my own. Thought of doing something on Thirukural discourse parsing but only then I found, I should have had a knowledge base to do a discourse parsing. So thought of automatically constructing a dictionary for Thirukural and submitted this for the research day in the year 2016 that is celebrated every year on Feb 28 at SRM. This won the gold medal in the faculty track for me:) and I could publish this in Applied Artificial Intelligence Journal published by Taylor and Francis publisher. . https://www.tandfonline.com/doi/citedby/10.1080/08839514.2018.1481590?scroll=top&needAccess=true
For the next research day in 2017 too, I wanted to participate but I wanted to do something for Kurunthogai. Since my school days, Kurunthogai stayed vivid in my mind for no reason. I started to read on that and that's when I understood that It is about tales of our ancestors lived during the Sangam age. I thought of an Information Extraction system which could extract information automatically about the flora, fauna , the types of vessels they used and food they ate from Kurunthogai. The UNL knowledge base and this excellent website on sangam literature, https://sangamtranslationsbyvaidehi.com/ is an awesome webpage helped a lot. Vaidehi madam has done an excellent job which could help more NLP researchers to dig more on Sangam literature. Ms. Nandhini Karky's podcast on Natrinai is also awesome https://nandinikarky.com and is a wonderful linguistic knowledge base for NLP researchers. Back to my research day , I won the gold medal for the second time again. This work is published in Sadhana journal published by Springer and can be downloaded right from here.
Kurunthogai paper
.pdf
Download PDF • 674KB
Soon after this, I got my PhD Guideship and out of 6 Phd Scholars, 4 have somehow incorporated Tamil in their research work and out of this 4 scholars, 1 Phd scholar's entire PhD is on Tamil literature.We tried discourse parsing on Thirukural as her first work and it did won the gold medal for us for the third consequent time in the year 2018:) We published this in https://ltrc.iiit.ac.in/icon2019/ but paper details are yet to be indexed. There are few more experiments on Tamil executed and published in Tamil Internet Conference 2019 can be downloaded from here.
proverbs
.pdf
Download PDF • 237KB
to prefer Tamil words
.pdf
Download PDF • 146KB
Availability of resources is still pulling us back. We still use our Anna University Morphological Analyser and we have also used Karky Research Foundation Analyser. We have to find more open source linguistic tools, bench mark data sets and yes most importantly, "Fund" from Government to support this. Thanks so much again, Ezhil and team(Build for Tamil twitter team ) for suggesting me to write this series :)Hope I continue adding more on this series in the years to come. Let us see how it goes:)
Thanks for reading
Subalalitha