Print Email Facebook Twitter Neural Network Output Optimization Using Interval Analysis Title Neural Network Output Optimization Using Interval Analysis Author De Weerdt, E. Chu, Q.P. Mulder, J.A. Faculty Aerospace Engineering Department Control & Operations Date 2009-04-03 Abstract The problem of output optimization within a specified input space of neural networks (NNs) with fixed weights is discussed in this paper. The problem is (highly) nonlinear when nonlinear activation functions are used. This global optimization problem is encountered in the reinforcement learning (RL) community. Interval analysis is applied to guarantee that all solutions are found to any degree of accuracy with guaranteed bounds. The major drawbacks of interval analysis, i.e., dependency effect and high-computational load, are both present for the problem of NN output optimization. Taylor models (TMs) are introduced to reduce these drawbacks. They have excellent convergence properties for small intervals. However, the dependency effect still remains and is even made worse when evaluating large input domains. As an alternative to TMs, a different form of polynomial inclusion functions, called the polynomial set (PS) method, is introduced. This new method has the property that the bounds on the network output are tighter or at least equal to those obtained through standard interval arithmetic (IA). Experiments show that the PS method outperforms the other methods for the NN output optimization problem. Subject feedforward neural networks (FFNNs)global optimizationinclusion functioninterval analysisoptimization methodspolynomial setradial basis function neural networks (RBFNNs)taylor expansiontaylor model (TM) To reference this document use: http://resolver.tudelft.nl/uuid:b0d2cc9f-c6b9-49b0-91e5-d5ef37e32ecb Publisher IEEE ISSN 1045-9227 Source IEEE Transactions on Neural Networks, 20 (4), 2009 Part of collection Institutional Repository Document type journal article Rights (c) 2009 The Author(s) ; IEEE Files PDF deWeerdt_2009.pdf 1020.63 KB Close viewer /islandora/object/uuid:b0d2cc9f-c6b9-49b0-91e5-d5ef37e32ecb/datastream/OBJ/view