search

kdmalc / personalization-privacy-risk

Privacy analysis for ML and classical filtering personalization parameters
0 stars 0 forks source link

Why does non-exponential updating scheme blow up? #10

Closed kdmalc closed 1 year ago

kdmalc commented 1 year ago

The "correct" code ought to be to simply add the original local threshold to the current threshold so that each update gets trained on the same number of iters. When I do this the error blows up. Instead, if I leave it such that each update doubles the previous number of iters, then the loss is somewhat contained, but doesn't really drop between updates and certainly not within updates

kdmalc commented 1 year ago

Idk, I just switched it back and it still happens for maroon user but doesn't seem to hurt anyone else

kdmalc commented 1 year ago

Superseded by PFL Non-IID codebase.