CV-梯度下降

GD θ n e x t = θ l a s t − ω f ′ ( θ i ) \Large \theta_{next} = \theta_{last} - \omega f'(\theta_i) θnext​=θlast​−ωf′(θi​) 直到 f ′ ( θ n e x t ) = 0 \Large f'(\theta_{next}) = 0 f′(θnext​)=0 对于一般的凸函数，其

>>阅读原文<<