The two well known learning algorithms of recurrent neural networks are the back-propagation (Rumelhart & el al., Werbos) and the forward propagation (Williams and Zipser). The main drawback of back-propagation is its off-line backward path in time for error cumulation. This violates the on-line requirement in many practical applications. Although the forward propagation algorithm can be used i...

In this paper, we propose a variable metric version of Tseng’s algorithm (the forward-backward-forward algorithm: FBF) combined with extrapolation from the past that includes error terms for finding zero sum maximally monotone operator and Lipschitzian in Hilbert spaces. The algorithm, which is modified forward-reflected-backword method, endowed metrics terms. Primal-dual algorithms are also pr...

the phase transformation phenomenon due to the crystallographic change of shape memory alloys subjected to mechanical or thermal loading is very complicated. regarding the thermo-mechanical coupling effects in shape memory alloys, in case of high  loading rates, heat generation/absorption during the forward/reverse transformation, will lead in temperature-dependent variation and consequently af...

to date, a number of numerical methods, including the popular finite-difference time domain (fdtd) technique, have been proposed to simulate ground-penetrating radar (gpr) responses. despite having a number of advantages, the finite-difference method also has pitfalls such as being very time consuming in simulating the most common case of media with high dielectric permittivity, causing the for...

Basic definitions concerning the multi-layer feed-forward neural networks are given. The back-propagation training algorithm is explained. Partial derivatives of the objective function with respect to the weight and threshold coefficients are derived. These derivatives are valuable for an adaptation process of the considered neural network. Training and generalisation of multi-layer feed-forwar...

A method is described for quickly inferring a player’s goals from low-level inputs in an action-adventure game. Fast Forward Bayesian Models (FFBMs) use the very efficient forward algorithm normally used in the Forward-Backward algorithm, but they use transition matrices and observation functions specific to the game’s current state. In experiments with both a veteran player and novice players ...

We present two efficient search algorithms for real-time spoken language systems. The first called the Word-Dependent N-Best algorithm is an improved algorithm for finding the top N sentence hypotheses. The new algorithm is shown to perform as well as the Exact Sentence-Dependent algorithm presented previously but with an order of magnitude less computation. The second algorithm is a fast match...