Polynomial expansions are important in the analysis of neural network nonlinearities. They have been applied thereto addressing well-known difficulties in verification, explainability, and security. Existing approaches span classical Taylor and Chebyshev methods, asymptotics, and many numerical approaches. We find that, while these have useful properties individually, such as exact error formulas, adjustable domain, and robustness to undefined derivatives, there are no approaches that provide a consistent method, yielding an expansion with all these properties. To address this, we develop an analytically modified integral transform expansion (AMITE), a novel expansion via integral transforms modified using derived criteria for convergence. We show the general expansion and then demonstrate an application for two popular activation functions: hyperbolic tangent and rectified linear units. Compared with existing expansions (i.e., Chebyshev, Taylor, and numerical) employed to this end, AMITE is the first to provide six previously mutually exclusive desired expansion properties, such as exact formulas for the coefficients and exact expansion errors. We demonstrate the effectiveness of AMITE in two case studies. First, a multivariate polynomial form is efficiently extracted from a single hidden layer black-box multilayer perceptron (MLP) to facilitate equivalence testing from noisy stimulus-response pairs. Second, a variety of feedforward neural network (FFNN) architectures having between three and seven layers are range bounded using Taylor models improved by the AMITE polynomials and error formulas. AMITE presents a new dimension of expansion methods suitable for the analysis/approximation of nonlinearities in neural networks, opening new directions and opportunities for the theoretical analysis and systematic testing of neural networks.

Classification of unexploded ordnance (UXO) in a fast and accurate manner is an imperative need in both the military and humanitarian demining contexts. An automated algorithm would significantly make the task safer, more convenient and less expensive. This paper targets this problem with the development of an image based approach and employs convolutional networks with images that were acquired on a custom camera setup. As access to multiple samples of ordnance can often be an issue, this work focuses on training the algorithm using a limited dataset and for optimizing its performance when ordnance is found in varying conditions. In particular, seven classes of UXO are considered with 420 images used for the training set and 140 images for the validation set. Using images that account for variation, a classifier was developed which was found to be 97.1% accurate.

This paper is concerned with predicting the occurrence of periventricular leukomalacia (PVL) using vital and blood gas data which are collected over a period of 12 h after the neonatal cardiac surgery. A data mining approach has been employed to generate a set of rules for classification of subjects as healthy or PVL affected. In view of the fact that blood gas and vital data have different sampling rates, in this study we have divided the data into two categories: 1) high resolution (vital), and 2) low resolution (blood gas), and designed a separate classifier based on each data category. The developed algorithm is composed of several stages; first, a feature pool has been extracted from each data category and the extracted features have been ranked based on the data reliability and their mutual information content with the output. An optimal feature subset with the highest discriminative capability has been formed using simultaneous maximization of the class separability measure and mutual information of a set. Two separate decision trees (DTs) have been developed for the classification purpose and more importantly to discover hidden relationships that exist among the data to help us better understand PVL pathophysiology. The DT result shows that high amplitude 20 min variations and low sample entropy in the vital data and the defined out of range index as well as maximum rate of change in blood gas data are important factors for PVL prediction. Low sample entropy represents lack of variability in hemodynamic measurement, and constant blood pressure with small fluctuations is an important indicator of PVL occurrence. Finally, using the different time frames of data collection, we show that the first 6 h of data contain sufficient information for PVL occurrence prediction.

Abstract The objective of this study is to develop an algorithm to detect and classify six types of electrocardiogram (ECG) signal beats including normal beats (N), atrial pre‐mature beats (A), right bundle branch block beats (R), left bundle branch block beats (L), paced beats (P), and pre‐mature ventricular contraction beats (PVC or V) using a neural network classifier. In order to prepare an appropriate input vector for the neural classifier several pre‐processing stages have been applied. Initially, a signal filtering method is used to remove the ECG signal baseline wandering. Continuous wavelet transform is then applied in order to extract features of the ECG signal. Next, principal component analysis is used to reduce the size of the data. A well‐known neural network architecture called the multi‐layered perceptron neural network is then utilized as the final classifier to classify each ECG beat as one of six groups of signals under study. Finally, the MIT‐BIH database is used to evaluate the proposed algorithm, resulting in 99.5% sensitivity, 99.66% positive predictive accuracy and 99.17% total accuracy.

A study is presented on the application of particle swarm optimization (PSO) for estimation of parameters in chaotic systems. The parameter estimation is formulated as a nonlinear optimization problem using PSO to minimize the synchronization error for the observable states of the actual system and its mathematical model. The procedure is illustrated using a typical chaotic system of Lorenz equations. The effectiveness of different variants of PSO on parameter estimation is studied with a wide search range of parameters. The results show the capability of the proposed PSO based approach in estimating the chaotic system parameters even in the presence of observation noise.

A study is presented on the application of particle swarm optimization (PSO) to design intelligent autopilots for ship steering. Two versions of PSO-conventional and anti-predatory (APSO) - have been used. The autopilot consists of a fuzzy logic controller (FLC) emulating the characteristics of manual ship steering. The parameters for the FLC are optimized using PSO and APSO. The robustness of the autopilot based on the optimized FLC is illustrated through the simulation of a realistic nonlinear ship model. The steering performance of the optimal FLC is compared with a PID type autopilot designed with and without PSO.

A new and efficient kinematic position and velocity solution scheme for spatial hyper-redundant manipulators is presented. The manipulator's arm has discrete links and universal joints. Backbone curve concepts and a modal approach are used to resolve the manipulator's redundancy. The effects of the mode shapes and the slope of the backbone curve at the starting point on the workspace are studied. It is shown that the usage of conventional mode shapes limits the workspace of the hyper-redundant arm. By introducing new mode shapes, an improved workspace is obtained. A simple and efficient recursive fitting method is introduced to avoid complications involved with solving systems of nonlinear algebraic equations. This method also guarantees the existence of solutions for the inverse kinematic problem at the velocity level. Velocity properties of the backbone curve are investigated and the inverse velocity propagation is solved for the spatial hyper-redundant arm. The velocity propagation scheme is recursive and is efficiently applicable to any number of links.

Biasing, once regarded as "cheating" in the machine learning community, is now understood and accepted as a necessary part of learning. However, despite its wide acceptance and recognition, biasing has never been studied as a separate research issue, except by Hailu & Sommer (1999), who made an attempt to shed light on the relationship between the quality of the bias and learning trials. So far, the general view held in biasing a learning system is to look for a bias that maximally collapses the search space. It is well-known, however, that reckless reduction of the search space often leads to sub-optimal learning. Regardless of the final level of optimality, this paper challenges this broadily accepted biasing scheme from the point of view of accelerating the learning process itself. Is a large search space a definitive indication of slow learning? We give a non-affirmative answer to this dogma by presenting a typical robot learning scenario. Experiments clearly indicate that, in spite of its large search space, a bias that is derived from the unique characteristics of the task shows better learning behavior than a bias that reduces the search space aggressively.