1: \begin{abstract}

2: %We take a different approach to analyze the generalized proximal point algorithm (GPPA). Our analysis justifies the structure imposed on the distance metric in establishing the convergence rate of GPPA.

3: We provide a new perspective to understand proximal gradient algorithms. We show that both proximal gradient algorithm (PGA) and Bregman proximal gradient algorithm (BPGA) can be viewed as generalized proximal point algorithm (GPPA), based on which more accurate convergence rates of PGA and BPGA are obtained directly. Furthermore, based on GPPA framework, we incorporate the back-tracking line search scheme into PGA and BPGA, and analyze the convergence rate with numerical verification.

4: %In this paper, we take a different approach to analyze a generalized proximal point algorithm~(GPPA), and recover the existing convergence rate of GPPA having a certain structure. Our analysis provides justification of imposing such a structure on the distance metric in establishing the convergence rate of GPPA. We further show that both proximal gradient algorithm~(PGA) and its generalized version---Bregman proximal gradient algorithm~(BPGA) for minimizing a sum of two convex functions, can be viewed as GPPA. Thus, this new perspective provides a direct way to obtain the convergence rate of PGA and BPGA based on that of GPPA, which significantly simplifies the existing analysis. Furthermore, the convergence rate obtained in this way is more accurate than the existing rate.

5: 

6: %Moreover, the GPPA scheme allows a gradient descent view of PGA via the envelope function of the composite objective.

7: \end{abstract}