Conversation

Notices

1. Embed this notice
   blaue_Fledermaus (blaue_fledermaus@mstdn.io)'s status on Wednesday, 06-Nov-2024 17:55:09 JST blaue_Fledermaus

   • Dana Fried

   @tess
   Genuinely curious, won't this just help train more robust models?
   Can't those training the models poison their datasets themselves and instruct the program to look for it?