Dynamics and its stability of Boltzmann-machine learning algorithm for gray scale image restoration J. Inoue (Hokkaido Univ.) and K. Tanaka (Tohoku Univ.) The 3 rd International Symposium on Slow Dynamics in Complex Systems in Sendai November 2003
Plan of this talk Bayesian image restoration and hyper-parameter estimation Boltzmann-machine learning algorithm for the hyper- parameter estimation Dynamic behavior of the BML algorithm Stability of the solution Concluding remarks
Bayesian image restoration OriginalCorrupted We treat images and the degrading process as spin systems
Definitions of the model by spin systems Original Corrupted : Hyper-parameters (true value)
Bayesian approach and MPM estimation takes its minimum at [Inoue and Carlucci (2001)]
Maximization of the marginal likelihood via Boltzmann-machine learning algorithm takes its maximum aton average We evaluate the data-averaged BML algorithm at the mean-field level : [Inoue and Tanaka (2003)]
Dynamic behavior of the hyper-parameters are integrated numerically
Analysis of the stability Expand the BML equations around and check the sign of eigenvalues of the Hessian A The solution is asymptotically stable
True hyper-parameter dependence of the stability The solution of the BML algorithm is asymptotically stable as long as the solution is identical to the true value of the hyper-parameters (fixed)
Behavior of the BML algorithm around the solution Trajectories in the hyper-parameter space (around the solution)
Concluding remarks We investigated dynamic behavior and its stability of the BML algorithm for gray scale image restoration We derived the data-averaged BML equations The solution is asymptotically stable as long as the solution is identical to the true value of the hyper- parameters More details of the present study are available at Send