Differentially Private Non-Convex Optimization under the KL Condition with Optimal Rates
arxiv(2023)
摘要
We study private empirical risk minimization (ERM) problem for losses
satisfying the (γ,κ)-Kurdyka-Łojasiewicz (KL) condition. The
Polyak-Łojasiewicz (PL) condition is a special case of this condition when
κ=2. Specifically, we study this problem under the constraint of ρ
zero-concentrated differential privacy (zCDP). When κ∈[1,2] and the
loss function is Lipschitz and smooth over a sufficiently large region, we
provide a new algorithm based on variance reduced gradient descent that
achieves the rate
Õ((√(d)/n√(ρ))^κ) on the
excess empirical risk, where n is the dataset size and d is the dimension.
We further show that this rate is nearly optimal. When κ≥ 2 and the
loss is instead Lipschitz and weakly convex, we show it is possible to achieve
the rate Õ((√(d)/n√(ρ))^κ)
with a private implementation of the proximal point method. When the KL
parameters are unknown, we provide a novel modification and analysis of the
noisy gradient descent algorithm and show that this algorithm achieves a rate
of
Õ((√(d)/n√(ρ))^2κ/4-κ)
adaptively, which is nearly optimal when κ = 2. We further show that,
without assuming the KL condition, the same gradient descent algorithm can
achieve fast convergence to a stationary point when the gradient stays
sufficiently large during the run of the algorithm. Specifically, we show that
this algorithm can approximate stationary points of Lipschitz, smooth (and
possibly nonconvex) objectives with rate as fast as
Õ(√(d)/n√(ρ)) and never worse than
Õ((√(d)/n√(ρ))^1/2). The latter
rate matches the best known rate for methods that do not rely on variance
reduction.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要