P
Initializing...
(A.14), p. 37 — the regret of Algorithm 1 is at most 𝔼 Σ_{ℓ≤L} (1 + V(S_ℓ))(R(S*, v) − R(S_ℓ, v)) · Prove2Me