Near-Optimal Complexity of Finite-Sum Nonconvex-Strongly-Concave Minimax Optimization

Under individual smoothness, known algorithms for nonconvex finite-sum optimization use $O(n+\sqrt nΔL_{\max}/\varepsilon^2)$ incremental first-order oracle (IFO) calls, while existing lower bounds miss the factor $\sqrt n$ in the second term. We prove the matching lower bound $Ω(n+\sqrt nΔL_{\max}/\varepsilon^2)$ for randomized IFO algorithms, allowing component indices and query points to depend on the complete preceding history. Thus PAGE and SPIDER are minimax optimal up to universal constants under individual and mean-squared smoothness. Under the global Polyak-Lojasiewicz (PL) condition, prior lower bounds did not capture the correct dependence on $n$ under individual smoothness, and the regime of small $κ$ remained open for randomized IFO algorithms. We prove an $Ω(n+κ_{\max}\sqrt n\log(Δ/\varepsilon))$ lower bound for $κ_{\max}\ge\sqrt n$, and an $Ω(n+n\log(Δ/\varepsilon)/(1+\log(\sqrt n/κ_{\max})))$ lower bound for $3\leκ_{\max}\le\sqrt n$. Restarted PAGE attains matching upper bounds under mean-squared smoothness in terms of $κ_{\rm ms}=L_{\rm ms}/μ$. Our lower bounds use dense weak hiding. Each hidden direction is spread across all components, making individual queries weakly informative while preserving the direction in the average. A sufficiently large dimension makes premature activation of future stages unlikely within the query budget. Balancing the cost of revealing a stage against the number of stages permitted by individual smoothness yields the missing $\sqrt n$ factor.

We establish complexity lower bounds for stochastic first-order algorithms in nonconvex--concave minimax optimization, allowing algorithms to use variance reduction. Our main contribution is a lower bound for a zero-respecting algorithm class that permits variance reduction, extending beyond the algorithmic restrictions imposed by some existing lower bounds. We consider objectives with an $L$-Lipschitz continuous joint gradient, a compact convex dual domain of Euclidean radius at most $D_Y$, and a primal value function, defined by maximizing the objective over the dual variable, with initial suboptimality at most $Δ$. The target accuracy $\varepsilon$ is measured by the gradient norm of the Moreau envelope of the constrained primal value function with parameter $1/(2L)$. Under an unbiased stochastic first-order oracle with variance at most $σ^2$ and mean-square smoothness, we prove the lower bound $Ω\!\left(L^2D_YΔ\varepsilon^{-3}+L^3D_Y^2Δσ^2\varepsilon^{-6}\right)$. This result quantifies the dependence on accuracy, dual-domain radius, and oracle noise even when variance reduction is allowed. We also establish complementary lower bounds for nonconvex--strongly-concave minimax optimization. With dual strong-concavity parameter $μ>0$ and condition number $κ:=L/μ$, we obtain $Ω\!\left(LΔ\sqrtκ\,\varepsilon^{-2}+LΔκσ^2\varepsilon^{-4}\right)$ under the bounded-variance oracle model. Under the additional mean-square smoothness condition with constant $\bar L$, we obtain $Ω\!\left(LΔ\sqrtκ\,\varepsilon^{-2}+Δ\bar Lσκ^{3/2}\varepsilon^{-3}\right)$. Together, these results identify complexity barriers across the concave and strongly concave regimes, with the main nonconvex--concave bound remaining valid for algorithms that use variance reduction.

We study whether the linear condition-number dependence in the stochastic complexity of SAPD+ is necessary for nonconvex-strongly-concave minimax optimization. For jointly $L$-smooth objectives with dual strong-concavity parameter $μ$, we prove a lower bound that matches the SAPD+ upper bound under the same Moreau-envelope stationarity criterion and the same primal-dual initialization gap. Specifically, when $σ\ge\varepsilon$, the worst-case complexity of zero-respecting algorithms is $Θ(κLGσ^2\varepsilon^{-4})$ in the stated accuracy regime, where $κ=L/μ$, $G$ bounds the initial primal-dual gap, and $σ^2$ bounds the variance of a general unbiased first-order oracle. The lower bound is realized on a smooth problem class with a bounded dual box. Our construction routes each link of a nonconvex zero-chain through a dual gradient of magnitude proportional to $\varepsilon/\sqrtκ$, while an undiscovered primal coordinate prevents stationarity. It also yields the primal-gradient lower bound $Ω(LΔ(\sqrtκ\varepsilon^{-2}+κσ^2\varepsilon^{-4}))$ after combination with the known deterministic bound, where $Δ$ bounds the initial primal function gap.