Identification Results in Neural Network Theory and Linear Operator Theory


Author / Producer

Date

2021-03

Publication Type

Doctoral Thesis

ETH Bibliography

yes

Citations

Altmetric

Data

Abstract

This thesis addresses two distinct problems in the theory of system identification by means of various novel techniques relying on complex analysis. Specifically, we study the analytic continuation of deep neural networks with meromorphic nonlinearities, leading to a full resolution of the question posed in (Fefferman, 1994) of neural network identifiability for the tanh nonlinearity, and use the theory of interpolation and sampling of entire functions to derive necessary and sufficient conditions for the identifiability of linear time-varying (LTV) systems characterized by a (possibly infinite) discrete set of delay-Doppler shifts.

Publication status

published

Editor

Contributors

Examiner : Bandeira, Afonso S.
Examiner : Fefferman, Charles

Book title

Journal / series

Volume

Pages / Article No.

Publisher

ETH Zurich

Event

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

system identification; complex analysis; neural networks; Linear time-varying system

Organisational unit

02140 - Dep. Inf.technologie und Elektrotechnik / Dep. of Inform.Technol. Electrical Eng.

Notes

Funding

Related publications and datasets