Chelsea Finn: Supervised ML methods (i.e. ERM) assume that train & test data are from the same distribution, & deteriorate when this assumption is broken.
To help, we introduce adaptive risk minimization (ARM):
With M Zhang, H Marklund @abhishekunique7 @svlevine
10 replies, 645 likes
Sergey Levine: In the real world, the test distribution never actually matches the training distribution. Adaptive risk minimization (ARM) addresses distributional shift by adapting to it, without labels -- just from seeing a group of test inputs rather than a single individual test input. https://t.co/vOrVzw4pDX
1 replies, 233 likes
Found on Jul 07 2020 at https://arxiv.org/pdf/2007.02931.pdf