Tool
WALNUT
This repository contains the baseline code for the paper published in NAACL 2022: “WALNUT: A Benchmark on Weakly Supervised Learning for Natural Language Understanding”. Detailed description about the data sets and methods can be manuscript…
Publication
Grounded Language-Image Pre-training
Project
Sign Language Inclusive Play
We explore the use of the Ludic Design methodology to address two critical needs of the deaf and hard of hearing (DHH) community. We explore the use of the Ludic Design methodology to address two…
Tool
ToxiGen
Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle…