DILL Lab 🌿

University of Southern CaliforniaViterbi CSUSC NLP

dill-canva.png

SAL 329

941 Bloom Walk

Los Angeles, CA 90089

DILL is a lab dedicated to the study of Data, Interpretability, Language and Learning, led by Swabha Swayamdipta. DILL is focused on automatically estimating the quality of datasets for models, efficient pretraining, and semi-automatically building datasets that help models learn better. We also emphasize on special aspects of language that affect dataset quality, such as subjectivity and ambiguity.

DILL celebrated Thanksgiving 2022 together. From right to left: Sayan, Jaspreet, Swabha and guest Ali Omrani with a very good host, Kili (front).

news

Jan 24, 2023 Friend of the lab, Eunice Jun gave an invited talk at the group meeting.
Jan 9, 2023 Software engineering PhD student Tooraj Helmi will be a guest at the DILL for the spring semester.
Nov 22, 2022 DILL celebrated Thanksgiving together with a couple of guests (Ali Omrani and Souti Chattopadhyay) at Swabha’s.
Nov 22, 2022 Sayan and Jaspreet presented a brief overview of their research projects at the USC-NLP lunch.
Nov 18, 2022 We attended the SoCalNLP Symposium where Sayan and Jaspreet presented their latest research posters, and Swabha gave an invited talk.
Aug 15, 2022 First day at USC for the entire DILL Lab!
Apr 15, 2022 We now have four new lab members!