Hi all, This study about the GPT-2 and GPT-3 machine learning models outputting data verbatim from the training data set has interesting copyright, licensing, source and privacy implications that could be interesting to take into account for the Debian machine learning policy: https://bair.berkeley.edu/blog/2020/12/20/lmmem/ https://news.ycombinator.com/item?id=25542011 PS: please CC me on any replies that you would like me to read. -- bye, pabs https://wiki.debian.org/PaulWise
Attachment:
signature.asc
Description: This is a digitally signed message part