The authors define functional Bregman divergences in reproducing kernel Hilbert spaces by composing Bregman generators with kernel mean embeddings, yielding an easily estimable generalization of MMD.
Proof.(2)⇒(3) is immediate, since∇Φ(f) =T f+b, and therefored Φ(f, g) = 1 2 ⟨f−g, T(f−g)⟩ F
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Generalising maximum mean discrepancy: kernelised functional Bregman divergences
The authors define functional Bregman divergences in reproducing kernel Hilbert spaces by composing Bregman generators with kernel mean embeddings, yielding an easily estimable generalization of MMD.