I am interested in how abstractness and grounding can be used to make language models more flexible when it comes to dialectal differences. I am a heritage speaker of Sri Lankan Tamil, a dialect that is in danger of disappearing due to half a century of persecution, so making systems that don’t further alienate speakers of uncommon dialects is very close to my heart.

Here are some slides on what I’m currently doing.

Anjali Kantharuban

email linkedin github twitter How do I get research? How do I get a teaching position? Should I double major?