A Modularization Scheme for Feedforward Networks
نویسنده
چکیده
This article proposes a modularization scheme for feedforward networks based on controllable internal representations. Control is achieved by replacing hidden units with pretrained modules that constrain internal patterns of activity to desired subsets. In the case of auto-associative feedforward networks these subsets can be seen as module interfaces. If enough a priori knowledge about a system is available, hierarchical systems with separately trainable and exchangable modules can be built.
منابع مشابه
Asynchronous Networks: Modularization of Dynamics Theorem
Building on the first part of this paper, we develop the theory of functional asynchronous networks. We show that a large class of functional asynchronous networks can be (uniquely) represented as feedforward networks connecting events or dynamical modules. For these networks we can give a complete description of the network function in terms of the function of the events comprising the network...
متن کاملThe pandemonium system of reflective agents
The Pandemonium system of reflective MINOS agents solves problems by automatic dynamic modularization of the input space. The agents contain feedforward neural networks which adapt using the backpropagation algorithm. We demonstrate the performance of Pandemonium on various categories of problems. These include learning continuous functions with discontinuities, separating two spirals, learning...
متن کاملFeedforward Neural Network Architectures for Complex Classi cation Problems
This paper presents two neural network design strategies for incorporating a pri-ori knowledge about a given problem into the feedforward neural networks. These strategies aim at obtaining tractability and reliability for solving complex classiica-tion problems by neural networks. The rst type strategy based on multistage scheme decomposes the problem into manageable ones for reducing the compl...
متن کاملAn Encoding Scheme for Cooperative Coevolutionary Feedforward Neural Networks
The cooperative coevolution paradigm decomposes a large problem into a set of subcomponents and solves them independently in order to collectively solve the large problem. This work introduces a novel encoding scheme for building subcomponents based on functional properties of a neuron. The encoding scheme is used for training feedforward neural networks. The results show that the proposed enco...
متن کاملStrategic Application of Feedforward Neural Networks to Large-Scale Classification
Feedforward neural networks have been successfully applied to a variety of classification problems, but th e number of classes used for experiments was too small to app ly the results directly to large-scale problems. This pap er presents several st rategies for applying feedforward neur al networks to large-scale, complex classificat ion problems: a two-stage classification scheme, a rapid lea...
متن کامل