A new method enhances language models by integrating knowledge across languages.
― 7 min read
Siyuan Wu
A study introduces SeqMIA to improve privacy against membership inference attacks.
― 6 min read
Examining the risks and strategies of model hijacking in federated learning systems.
― 5 min read