Meyer-Bäse, Anke ; Gruber, Peter ; Foo, Simon ; Theis, Fabian J.
Alternative Links zum Volltext:DOI
Zusammenfassung
This contribution describes a previous termneural network that selfnext term-organizes to recover the underlying original previous termsourcesnext term from typical sensor signals. No particular information is required about the statistical properties of the previous termsourcesnext term and the coefficients of the linear transformation, except the fact that the previous termsourcenext term ...
Zusammenfassung
This contribution describes a previous termneural network that selfnext term-organizes to recover the underlying original previous termsourcesnext term from typical sensor signals. No particular information is required about the statistical properties of the previous termsourcesnext term and the coefficients of the linear transformation, except the fact that the previous termsourcenext term signals are statistically independent and nonstationary. This is often true for real life applications. We propose an online learning solution using a previous termneural networknext term and use the nonstationarity of the previous termsourcesnext term to achieve the previous termseparation.next term The learning rule for the previous termnetwork'snext term parameters is derived from the steepest descent minimization of a time-dependent cost function that takes the minimum only when the previous termnetworknext term outputs are uncorrelated with each other. In this process divide the problem into two learning problems one of which is solved by an anti-Hebbian learning and the other by an Hebbian learning process. We also compare the performance of our algorithm with other solutions to this task.