Adaptive projected subgradient method

From Wikipedia, the free encyclopedia
Jump to: navigation, search

The adaptive projected subgradient method (APSM) is an algorithm, the goal of which is to minimize iteratively a sequence of cost functions.[1][2][3]

This algorithmic "tool" is general and has been used in several tasks, such as: online/adaptive parameter estimation, online classification,[4] adaptive distributed learning,[5] just to name a few.


  1. ^ Yamada, I.; Ogura, N. (2003). "Adaptive projected subgradient method and its applications to set theoretic adaptive filtering". "The Thirty-seventh Asilomar Conference on Signals, Systems & Computers, 2003". p. 600. doi:10.1109/ACSSC.2003.1291982. ISBN 0-7803-8104-1.  edit
  2. ^ Yamada, I.; Ogura, N. (2005). "Adaptive Projected Subgradient Method for Asymptotic Minimization of Sequence of Nonnegative Convex Functions". Numerical Functional Analysis and Optimization 25 (7–8): 593. doi:10.1081/NFA-200045806.  edit
  3. ^ Slavakis, Konstantinos; Yamada, Isao (5 August 2011). "The adaptive projected subgradient method constrained by families of quasi-nonexpansive mappings and its application to online learning". Ithaca, New York: Cornell University.
  4. ^ Slavakis, Konstantinos, Sergios Theodoridis, and Isao Yamada. "Online kernel-based classification using adaptive projection algorithms." Signal Processing, IEEE Transactions on 56.7 (2008): 2781-2796.
  5. ^ Chouvardas, Symeon, Konstantinos Slavakis, and Sergios Theodoridis. "Adaptive robust distributed learning in diffusion sensor networks." Signal Processing, IEEE Transactions on 59.10 (2011): 4692-4707.