Repository navigation

feat: the telescoping of the mix losses of exponential weights, stated once - #376

Merged

lmlexplore-automation[bot] merged 3 commits into

LeanMachineLearning:mainfrom

RemyDegenne:roadmap/anytime-exponential-weights-worker1

Nov 14, 2026

Merged

feat: the telescoping of the mix losses of exponential weights, stated once#376
lmlexplore-automation[bot] merged 3 commits into
LeanMachineLearning:mainfrom
RemyDegenne:roadmap/anytime-exponential-weights-worker1

Commits on Nov 14, 2026
- feat: the telescoping of the mix losses of exponential weights, stated once
  
  RemyDegenne
  and
  web-flow
  committed
  7554365
- doc: describe the mixability gap as the maximum of the FTRL stability term
  
  RemyDegenne
  and
  web-flow
  committed
  72dbf99
- refactor: state the mix loss inequalities as iff lemmas, used by AdaHedge
  
  RemyDegenne
  and
  web-flow
  committed
  43b6899