Litcius/Paper detail

Distributed Support Vector Machines Over Dynamic Balanced Directed Networks

Mohammadreza Doostmohammadian, Alireza Aghasi, Themistoklis Charalambous, Usman A. Khan

2021IEEE Control Systems Letters39 citationsDOIOpen Access PDF

Abstract

In this letter, we consider the binary classification problem via distributed Support Vector Machines (SVMs), where the idea is to train a network of agents, with limited share of data, to cooperatively learn the SVM classifier for the global database. Agents only share processed information regarding the classifier parameters and the gradient of the local loss functions instead of their raw data. In contrast to the existing work, we propose a continuous-time algorithm that incorporates network topology changes in discrete jumps. This hybrid nature allows us to remove chattering that arises because of the discretization of the underlying CT process. We show that the proposed algorithm converges to the SVM classifier over time-varying weight balanced directed graphs by using arguments from the matrix perturbation theory.

Topics & Concepts

Support vector machineDiscretizationComputer scienceClassifier (UML)Artificial intelligenceBinary numberBinary classificationData miningMachine learningAlgorithmPattern recognition (psychology)MathematicsArithmeticMathematical analysisDistributed Control Multi-Agent SystemsNeural Networks Stability and SynchronizationAdvanced Memory and Neural Computing
Distributed Support Vector Machines Over Dynamic Balanced Directed Networks | Litcius