Dynamical System Implementations of Sparse Bayesian Learning (bibtex)

by M. O'Shaughnessy, M. Davenport and C. Rozell

Abstract:

Despite its state of the art performance in many applications, the sparse Bayesian learning (SBL) procedure can be expensive to implement, limiting its use in practice. In this paper, we use the locally competitive algorithm (LCA) framework to develop two continuous time dynamical systems whose trajectories converge to a minimum of the SBL objective.T he resulting systems are neurally feasible and can be implemented using primitives from analog electronics, potentially opening the SBL procedure to new applications.

Reference:

Dynamical System Implementations of Sparse Bayesian LearningM. O'Shaughnessy, M. Davenport and C. Rozell. December 2019.

Bibtex Entry:

@CONFERENCE{oshaughnessy.19c,
       author = 	 {O'Shaughnessy, M. and Davenport, M.  and Rozell, C. },
       title = 	 {Dynamical System Implementations of Sparse Bayesian Learning},
      booktitle =	 {International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)},
      year =	 2019,
 	 address = {Guadeloupe, West Indies},
 	 month = dec,
 	 abstract = {Despite its state of the art performance in many applications, the sparse Bayesian learning (SBL) procedure can be expensive to implement, limiting its use in practice.  In this paper, we use the locally competitive algorithm (LCA) framework to  develop  two  continuous  time  dynamical  systems  whose trajectories  converge  to  a  minimum  of  the  SBL  objective.T he resulting systems are neurally feasible and can be implemented using primitives from analog electronics, potentially opening the SBL procedure to new applications.}
   }