Dynamic Cell Structures
نویسندگان
چکیده
Dynamic Cell Structures (DCS) represent a family of artificial neural architectures suited both for unsupervised and supervised learning. They belong to the recently [Martinetz94] introduced class of Topology Representing Networks (TRN) which build perlectly topology preserving feature maps. DCS empI'oy a modified Kohonen learning rule in conjunction with competitive Hebbian learning. The Kohonen type learning rule serves to adjust the synaptic weight vectors while Hebbian learning establishes a dynamic lateral connection structure between the units reflecting the topology of the feature manifold. In case of supervised learning, i.e. function approximation, each neural unit implements a Radial Basis Function, and an additional layer of linear output units adjusts according to a delta-rule. DCS is the first RBF-based approximation scheme attempting to concurrently learn and utilize a perfectly topology preserving map for improved performance. Simulations on a selection of CMU-Benchmarks indicate that the DCS idea applied to the Growing Cell Structure algorithm [Fritzke93] leads to an efficient and elegant algorithm that can beat conventional models on similar tasks.
منابع مشابه
Crashworthiness design of multi-cell tapered tubes using response surface methodology
In this article, crashworthiness performance and crushing behavior of tapered structures with four internal reinforcing plates under axial and oblique dynamic loadings have been investigated. These structures have a tapered form with five cross sections of square, hexagonal, octagonal, decagon and circular shape. In the first step, finite element simulations performed in LS-DYNA were validated ...
متن کاملApproximate Dynamic Analysis of Structures for Earthquake Loading Using FWT
Approximate dynamic analysis of structures is achieved by fast wavelet transform (FWT). The loads are considered as time history earthquake loads. To reduce the computational work, FWT is used by which the number of points in the earthquake record are reduced. For this purpose, the theory of wavelets together with filter banks are used. The low and high pass filters are used for the decompositi...
متن کاملNonlinear Analysis of Truss Structures Using Dynamic Relaxation (RESEARCH NOTE)
This paper presents a new approach for large-deflection analysis of truss structures employing the Dynamic Relaxation method (DR). The typical formulation for DR has been established utilizing the finite difference technique which is categorized as an explicit method. The special characteristic of the explicit method is its simple algebraic relationships in comparison with complicated matrix op...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1994