✦ LIBER ✦

Invariance priors for Bayesian feed-forward neural networks

✍ Scribed by Udo v. Toussaint; Silvio Gori; Volker Dose

Publisher: Elsevier Science
Year: 2006
Tongue: English
Weight: 1001 KB
Volume: 19
Category: Article
ISSN: 0893-6080
DOI: 10.1016/j.neunet.2006.01.017

No coin nor oath required. For personal study only.

✦ Synopsis

Neural networks (NN) are famous for their advantageous flexibility for problems when there is insufficient knowledge to set up a proper model. On the other hand, this flexibility can cause overfitting and can hamper the generalization of neural networks. Many approaches to regularizing NN have been suggested but most of them are based on ad hoc arguments. Employing the principle of transformation invariance, we derive a general prior in accordance with the Bayesian probability theory for feed-forward networks. An optimal network is determined by Bayesian model comparison, verifying the applicability of this approach. Additionally the prior presented affords cell pruning.

📜 SIMILAR VOLUMES

Behaviour of feed-forward neural network

Behaviour of feed-forward neural networks in invariant track finding

✍ A. Lanza; P. Vitulo; C. Cattaneo; M. Caresana; F. Panetsos 📂 Article 📅 1994 🏛 Elsevier Science 🌐 English ⚖ 719 KB

We performed several simulations with feed-forward neural networks using an idealized tracking apparatus with tracks invariant under translation and roto-translation transformations. Input information was provided to the networks without any preprocessing. We implemented 2 and 3 layer architectures

Feed-forward neural networks for seconda

Feed-forward neural networks for secondary structure prediction

✍ T.W. Barlow 📂 Article 📅 1995 🏛 Elsevier Science 🌐 English ⚖ 744 KB

A new scheme for training feed-forward n

A new scheme for training feed-forward neural networks

✍ Osama Abdel-Wahhab; M.A. Sid-Ahmed 📂 Article 📅 1997 🏛 Elsevier Science 🌐 English ⚖ 410 KB

In this paper we present a new algorithm, which is orders of magnitude faster than the delta rule, for training feed-forward neural networks. It provides a substantial improvement over the method of Scalero and Tepedelenlioglu (IEEE Trans. Signal Process. 40(1) (1992)) in both training time and nume

Feed forward neural networks modeling fo

Feed forward neural networks modeling for K–P interactions

✍ M.Y. El-Bakry 📂 Article 📅 2003 🏛 Elsevier Science 🌐 English ⚖ 102 KB

Artificial intelligence techniques involving neural networks became vital modeling tools where model dynamics are difficult to track with conventional techniques. The paper make use of the feed forward neural networks (FFNN) to model the charged multiplicity distribution of K-P interactions at high

5023833 Feed forward neural network for

5023833 Feed forward neural network for unary associative memory

📂 Article 📅 1993 🏛 Elsevier Science 🌐 English ⚖ 98 KB

Constructing and training feed-forward n

Constructing and training feed-forward neural networks for pattern classification

✍ Xudong Jiang; Alvin Harvey Kam Siew Wah 📂 Article 📅 2003 🏛 Elsevier Science 🌐 English ⚖ 329 KB