Globally optimal on-line learning rules for multi-layer neural networks

We present a method for determining the globally optimal on-line learning rule for a soft committee machine under a statistical mechanics framework. This rule maximizes the total reduction in generalization error over the whole learning process. A simple example demonstrates that the locally optimal rule, which maximizes the rate of decrease in generalization error, may perform poorly in comparison.

Efficient THz radiation from a nanocrystalline silicon-based multi-layer photomixer

In this paper we propose a novel type of multiple-layer photomixer based on amorphous/nano-crystalline-Si. Such a device implies that it could be possible to enhance the conversion efficiency from optical power to THz emission by increasing the absorption length and by reducing the device overheating through the use of substrates with higher thermal conductivity compared to GaAs. Our calculations show that the output power from a two-layer Si-based photomixer is at least ten times higher than that from conventional LT-GaAs photomixers at 1 THz.

Physical characteristics of localized surface plasmons resulting from nano-scale structured multi-layer thin films deposited on D-shaped optical fiber

Novel surface plasmonic optical fiber sensors have been fabricated using multiple coatings deposited on a lapped section of a single mode fiber. UV laser irradiation processing with a phase mask produces a nano-scaled surface relief grating structure resembling nano-wires. The resulting individual corrugations produced by material compaction are approximately 20 μm long with an average width at half maximum of 100 nm and generate localized surface plasmons. Experimental data are presented that show changes in the spectral characteristics after UV processing, coupled with an overall increase in the sensitivity of the devices to surrounding refractive index. Evidence is presented that there is an optimum UV dosage (48 joules) over which no significant additional optical change is observed. The devices are characterized with regards to change in refractive index, where significantly high spectral sensitivities in the aqueous index regime are found, ranging up to 4000 nm/RIU for wavelength and 800 dB/RIU for intensity. © 2013 Optical Society of America.

On the k-anonymization of time-varying and multi-layer social graphs

The popularity of online social media platforms provides an unprecedented opportunity to study real-world complex networks of interactions. However, releasing this data to researchers and the public comes at the cost of potentially exposing private and sensitive user information. It has been shown that a naive anonymization of a network by removing the identity of the nodes is not sufficient to preserve users’ privacy. In order to deal with malicious attacks, k -anonymity solutions have been proposed to partially obfuscate topological information that can be used to infer nodes’ identity. In this paper, we study the problem of ensuring k anonymity in time-varying graphs, i.e., graphs with a structure that changes over time, and multi-layer graphs, i.e., graphs with multiple types of links. More specifically, we examine the case in which the attacker has access to the degree of the nodes. The goal is to generate a new graph where, given the degree of a node in each (temporal) layer of the graph, such a node remains indistinguishable from other k-1 nodes in the graph. In order to achieve this, we find the optimal partitioning of the graph nodes such that the cost of anonymizing the degree information within each group is minimum. We show that this reduces to a special case of a Generalized Assignment Problem, and we propose a simple yet effective algorithm to solve it. Finally, we introduce an iterated linear programming approach to enforce the realizability of the anonymized degree sequences. The efficacy of the method is assessed through an extensive set of experiments on synthetic and real-world graphs.

Personalized recommendation on multi-layer context graph

Recommender systems have been successfully dealing with the problem of information overload. A considerable amount of research has been conducted on recommender systems, but most existing approaches only focus on user and item dimensions and neglect any additional contextual information, such as time and location. In this paper, we propose a Multi-Layer Context Graph (MLCG) model which incorporates a variety of contextual information into a recommendation process and models the interactions between users and items for better recommendation. Moreover, we provide a new ranking algorithm based on Personalized PageRank for recommendation in MLCG, which captures users' preferences and current situations. The experiments on two real-world datasets demonstrate the effectiveness of our approach.

Multi-layer attribute selection and classification algorithm for the diagnosis of cardiac autonomic neuropathy based on HRV attributes

Cardiac autonomic neuropathy (CAN) poses an important clinical problem, which often remains undetected due difficulty of conducting the current tests and their lack of sensitivity. CAN has been associated with growth in the risk of unexpected death in cardiac patients with diabetes mellitus. Heart rate variability (HRV) attributes have been actively investigated, since they are important for diagnostics in diabetes, Parkinson's disease, cardiac and renal disease. Due to the adverse effects of CAN it is important to obtain a robust and highly accurate diagnostic tool for identification of early CAN, when treatment has the best outcome. Use of HRV attributes to enhance the effectiveness of diagnosis of CAN progression may provide such a tool. In the present paper we propose a new machine learning algorithm, the Multi-Layer Attribute Selection and Classification (MLASC), for the diagnosis of CAN progression based on HRV attributes. It incorporates our new automated attribute selection procedure, Double Wrapper Subset Evaluator with Particle Swarm Optimization (DWSE-PSO). We present the results of experiments, which compare MLASC with other simpler versions and counterpart methods. The experiments used our large and well-known diabetes complications database. The results of experiments demonstrate that MLASC has significantly outperformed other simpler techniques.

Reducing electrowetting-on-dielectric actuation voltage using a novel electrode shape and a multi-layer dielectric coating

This paper presents design and fabrication of an electrowetting-on-dielectric (EWOD) device using a novel electrode shape and a multi-layer dielectric coating that reduce the actuation voltage of the device to less than 12.6 V. The fabrication of the EWOD electrodes is carried out in several steps including laser exposure, wet developing, etching, and stripping. A high-dielectric-constant multi-layer dielectric coating containing a 770 nm thick Polyvinylidene difluoride (PVDF) layer and a 1 µm thick Cyanoethyl pullulan (CEP) layer, is deposited on the EWOD electrodes for insulation. This multi-layer dielectric structure exhibits a high capacitance per unit area, and the novel electrode shape changes the actuation force at the droplet contact line reducing the voltage required to operate the device. In addition, an overlaying Teflon layer of 50 nm is placed on top of the dielectric structure to provide a hydrophobic surface for droplet manipulation. It is observed from the experiments that the electrode shape and the dielectric structure have contributed to the reduction of the actuation voltage of the EWOD device.

Characterization of Failure Modes in CFRP Composites - An ANN Approach

Carbon fiber reinforced polymer (CFRP) composite specimens with different thickness, geometry, and stacking sequences were subjected to fatigue spectrum loading in stages. Another set of specimens was subjected to static compression load. On-line acoustic Emission (AE) monitoring was carried out during these tests. Two artificial neural networks, Kohonen-self organizing feature map (KSOM), and multi-layer perceptron (MLP) have been developed for AE signal analysis. AE signals from specimens were clustered using the unsupervised learning KSOM. These clusters were correlated to the failure modes using available a priori information such as AE signal amplitude distributions, time of occurrence of signals, ultrasonic imaging, design of the laminates (stacking sequences, orientation of fibers), and AE parametric plots. Thereafter, AE signals generated from the rest of the specimens were classified by supervised learning MLP. The network developed is made suitable for on-line monitoring of AE signals in the presence of noise, which can be used for detection and identification of failure modes and their growth. The results indicate that the characteristics of AE signals from different failure modes in CFRP remain largely unaffected by the type of load, fiber orientation, and stacking sequences, they being representatives of the type of failure phenomena. The type of loading can have effect only on the extent of damage allowed before the specimens fail and hence on the number of AE signals during the test. The artificial neural networks (ANN) developed and the methods and procedures adopted show significant success in AE signal characterization under noisy environment (detection and identification of failure modes and their growth).

Quantification of impact of a cricket ball

The cricket is one of most popular games in the Asian subcontinent and its popularity is increasing every day. The issue of replacement of the cricket ball amidst the matches is always an uncomfortable situation for teams, umpires and even supporters. At present the basis of the replacement is solely on the judgement, experience and expertise of the umpires, which is subjective, controversial and debatable. In this paper, we have attempted a new approach to quantify the number of impacts or impact factor of a 4-piece leather ball used in the Intemational one-day and test cricket matches. This gives a more objective and scientific basis/ criteria for the replacement of the ball. Here, we have used a well known and widely used Thermal Infra-Red (TIR) imaging to capture the dynamics of the thermal profice of the cricket ball, which has been heated for about 15 seconds. The idea behind this approach is the simple observation that an old ball (ball with a few impacts) has different thermal signature/profice compared to the that of a new ball. This could be due to the change in the surface profice and internal structure, minor de-shaping, opening of seam etc. The TIR video and its frames, which is inherently noisy, are restored using Hebbian learning based FIR (sic), which performs optimal smoothing in relatively less number of iteration. We have focussed on the hottest region of the ball i.e., the inner core and tracked its thermal profice dynamics. Finally we have used multi layer perceptron model (MLP) to quantify the impact factor with fairly good accuracy.

An ANN based approach for voltage stability assessment

Artificial Neural Networks (ANNs) have recently been proposed as an alterative method for salving certain traditional problems in power systems where conventional techniques have not achieved the desired speed, accuracy or efficiency. This paper presents application of ANN where the aim is to achieve fast voltage stability margin assessment of power network in an energy control centre (ECC), with reduced number of appropriate inputs. L-index has been used for assessing voltage stability margin. Investigations are carried out on the influence of information encompassed in input vector and target out put vector, on the learning time and test performance of multi layer perceptron (MLP) based ANN model. LP based algorithm for voltage stability improvement, is used for generating meaningful training patterns in the normal operating range of the system. From the generated set of training patterns, appropriate training patterns are selected based on statistical correlation process, sensitivity matrix approach, contingency ranking approach and concentric relaxation method. Simulation results on a 24 bus EHV system, 30 bus modified IEEE system, and a 82 bus Indian power network are presented for illustration purposes.

Modelling nonlinear vehicle dynamics with neural networks

The nonlinear modelling ability of neural networks has been widely recognised as an effective tool to identify and control dynamic systems, with applications including nonlinear vehicle dynamics which this paper focuses on using multi-layer perceptron networks. Existing neural network literature does not detail some of the factors which effect neural network nonlinear modelling ability. This paper investigates into and concludes on required network size, structure and initial weights, considering results for networks of converged weights. The paper also presents an online training method and an error measure representing the network's parallel modelling ability over a range of operating conditions. Copyright © 2010 Inderscience Enterprises Ltd.