Improving short utterance i-vector speaker verification using utterance variance modelling and compensation techniques

This paper proposes techniques to improve the performance of i-vector based speaker verification systems when only short utterances are available. Short-length utterance i-vectors vary with speaker, session variations, and the phonetic content of the utterance. Well established methods such as linear discriminant analysis (LDA), source-normalized LDA (SN-LDA) and within-class covariance normalisation (WCCN) exist for compensating the session variation but we have identified the variability introduced by phonetic content due to utterance variation as an additional source of degradation when short-duration utterances are used. To compensate for utterance variations in short i-vector speaker verification systems using cosine similarity scoring (CSS), we have introduced a short utterance variance normalization (SUVN) technique and a short utterance variance (SUV) modelling approach at the i-vector feature level. A combination of SUVN with LDA and SN-LDA is proposed to compensate the session and utterance variations and is shown to provide improvement in performance over the traditional approach of using LDA and/or SN-LDA followed by WCCN. An alternative approach is also introduced using probabilistic linear discriminant analysis (PLDA) approach to directly model the SUV. The combination of SUVN, LDA and SN-LDA followed by SUV PLDA modelling provides an improvement over the baseline PLDA approach. We also show that for this combination of techniques, the utterance variation information needs to be artificially added to full-length i-vectors for PLDA modelling.

Modelling for the electricity distribution network

Global awareness for cleaner and renewable energy is transforming the electricity sector at many levels. New technologies are being increasingly integrated into the electricity grid at high, medium and low voltage levels, new taxes on carbon emissions are being introduced and individuals can now produce electricity, mainly through rooftop photovoltaic (PV) systems. While leading to improvements, these changes also introduce challenges, and a question that often rises is ‘how can we manage this constantly evolving grid?’ The Queensland Government and Ergon Energy, one of the two Queensland distribution companies, have partnered with some Australian and German universities on a project to answer this question in a holistic manner. The project investigates the impact the integration of renewables and other new technologies has on the physical structure of the grid, and how this evolving system can be managed in a sustainable and economical manner. To aid understanding of what the future might bring, a software platform has been developed that integrates two modelling techniques: agent-based modelling (ABM) to capture the characteristics of the different system units accurately and dynamically, and particle swarm optimization (PSO) to find the most economical mix of network extension and integration of distributed generation over long periods of time. Using data from Ergon Energy, two types of networks (3 phase, and Single Wired Earth Return or SWER) have been modelled; three-phase networks are usually used in dense networks such as urban areas, while SWER networks are widely used in rural Queensland. Simulations can be performed on these networks to identify the required upgrades, following a three-step process: a) what is already in place and how it performs under current and future loads, b) what can be done to manage it and plan the future grid and c) how these upgrades/new installations will perform over time. The number of small-scale distributed generators, e.g. PV and battery, is now sufficient (and expected to increase) to impact the operation of the grid, which in turn needs to be considered by the distribution network manager when planning for upgrades and/or installations to stay within regulatory limits. Different scenarios can be simulated, with different levels of distributed generation, in-place as well as expected, so that a large number of options can be assessed (Step a). Once the location, sizing and timing of assets upgrade and/or installation are found using optimisation techniques (Step b), it is possible to assess the adequacy of their daily performance using agent-based modelling (Step c). One distinguishing feature of this software is that it is possible to analyse a whole area at once, while still having a tailored solution for each of the sub-areas. To illustrate this, using the impact of battery and PV can have on the two types of networks mentioned above, three design conditions can be identified (amongst others): · Urban conditions o Feeders that have a low take-up of solar generators, may benefit from adding solar panels o Feeders that need voltage support at specific times, may be assisted by installing batteries · Rural conditions - SWER network o Feeders that need voltage support as well as peak lopping may benefit from both battery and solar panel installations. This small example demonstrates that no single solution can be applied across all three areas, and there is a need to be selective in which one is applied to each branch of the network. This is currently the function of the engineer who can define various scenarios against a configuration, test them and iterate towards an appropriate solution. Future work will focus on increasing the level of automation in identifying areas where particular solutions are applicable.

Blast response of reinforced concrete pile using fully coupled computer simulation techniques

Pile foundations transfer loads from superstructures to stronger sub soil. Their strength and stability can hence affect structural safety. This paper treats the response of reinforced concrete pile in saturated sand to a buried explosion. Fully coupled computer simulation techniques are used together with five different material models. Influence of reinforcement on pile response is investigated and important safety parameters of horizontal deformations and tensile stresses in the pile are evaluated. Results indicate that adequate longitudinal reinforcement and proper detailing of transverse reinforcement can reduce pile damage. Present findings can serve as a benchmark reference for future analysis and design.

Transforming morning to afternoon using linear regression techniques

Visual localization in outdoor environments is often hampered by the natural variation in appearance caused by such things as weather phenomena, diurnal fluctuations in lighting, and seasonal changes. Such changes are global across an environment and, in the case of global light changes and seasonal variation, the change in appearance occurs in a regular, cyclic manner. Visual localization could be greatly improved if it were possible to predict the appearance of a particular location at a particular time, based on the appearance of the location in the past and knowledge of the nature of appearance change over time. In this paper, we investigate whether global appearance changes in an environment can be learned sufficiently to improve visual localization performance. We use time of day as a test case, and generate transformations between morning and afternoon using sample images from a training set. We demonstrate the learned transformation can be generalized from training data and show the resulting visual localization on a test set is improved relative to raw image comparison. The improvement in localization remains when the area is revisited several weeks later.

A modified particle swarm optimization for correlated phenomena

The wide applicability of correlation analysis inspired the development of this paper. In this paper, a new correlated modified particle swarm optimization (COM-PSO) is developed. The Correlation Adjustment algorithm is proposed to recover the correlation between the considered variables of all particles at each of iterations. It is shown that the best solution, the mean and standard deviation of the solutions over the multiple runs as well as the convergence speed were improved when the correlation between the variables was increased. However, for some rotated benchmark function, the contrary results are obtained. Moreover, the best solution, the mean and standard deviation of the solutions are improved when the number of correlated variables of the benchmark functions is increased. The results of simulations and convergence performance are compared with the original PSO. The improvement of results, the convergence speed, and the ability to simulate the correlated phenomena by the proposed COM-PSO are discussed by the experimental results.

An efficient hybrid evolutionary optimization algorithm for daily volt/var control at distribution system including DGs

This paper presents a new hybrid evolutionary algorithm based on Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) for daily Volt/Var control in distribution system including Distributed Generators (DGs). Due to the small X/R ratio and radial configuration of distribution systems, DGs have much impact on this problem. Since DGs are independent power producers or private ownership, a price based methodology is proposed as a proper signal to encourage owners of DGs in active power generation. Generally, the daily Volt/Var control is a nonlinear optimization problem. Therefore, an efficient hybrid evolutionary method based on Particle Swarm Optimization and Ant Colony Optimization (ACO), called HPSO, is proposed to determine the active power values of DGs, reactive power values of capacitors and tap positions of transformers for the next day. The feasibility of the proposed algorithm is demonstrated and compared with methods based on the original PSO, ACO and GA algorithms on IEEE 34-bus distribution feeder.

A novel algorithm based on honey bee mating optimization for distribution harmonic state estimation including distributed generators

This paper presents a new algorithm based on honey-bee mating optimization (HBMO) to estimate harmonic state variables in distribution networks including distributed generators (DGs). The proposed algorithm performs estimation for both amplitude and phase of each harmonics by minimizing the error between the measured values from phasor measurement units (PMUs) and the values computed from the estimated parameters during the estimation process. Simulation results on two distribution test system are presented to demonstrate that the speed and accuracy of proposed distribution harmonic state estimation (DHSE) algorithm is extremely effective and efficient in comparison with the conventional algorithms such as weight least square (WLS), genetic algorithm (GA) and tabu search (TS).

Distribution feeder reconfiguration for loss minimization based on modified honey bee mating optimization algorithm

This paper presents an efficient algorithm for multi-objective distribution feeder reconfiguration based on Modified Honey Bee Mating Optimization (MHBMO) approach. The main objective of the Distribution feeder reconfiguration (DFR) is to minimize the real power loss, deviation of the nodes’ voltage. Because of the fact that the objectives are different and no commensurable, it is difficult to solve the problem by conventional approaches that may optimize a single objective. So the metahuristic algorithm has been applied to this problem. This paper describes the full algorithm to Objective functions paid, The results of simulations on a 32 bus distribution system is given and shown high accuracy and optimize the proposed algorithm in power loss minimization.

Optimization of an integrated model for automatic reduction and expansion of long queries

A long query provides more useful hints for searching relevant documents, but it is likely to introduce noise which affects retrieval performance. In order to smooth such adverse effect, it is important to reduce noisy terms, introduce and boost additional relevant terms. This paper presents a comprehensive framework, called Aspect Hidden Markov Model (AHMM), which integrates query reduction and expansion, for retrieval with long queries. It optimizes the probability distribution of query terms by utilizing intra-query term dependencies as well as the relationships between query terms and words observed in relevance feedback documents. Empirical evaluation on three large-scale TREC collections demonstrates that our approach, which is automatic, achieves salient improvements over various strong baselines, and also reaches a comparable performance to a state of the art method based on user’s interactive query term reduction and expansion.

Mechanism based emulation of dynamic simulation models : concept and application in hydrology

Many model-based investigation techniques, such as sensitivity analysis, optimization, and statistical inference, require a large number of model evaluations to be performed at different input and/or parameter values. This limits the application of these techniques to models that can be implemented in computationally efficient computer codes. Emulators, by providing efficient interpolation between outputs of deterministic simulation models, can considerably extend the field of applicability of such computationally demanding techniques. So far, the dominant techniques for developing emulators have been priors in the form of Gaussian stochastic processes (GASP) that were conditioned with a design data set of inputs and corresponding model outputs. In the context of dynamic models, this approach has two essential disadvantages: (i) these emulators do not consider our knowledge of the structure of the model, and (ii) they run into numerical difficulties if there are a large number of closely spaced input points as is often the case in the time dimension of dynamic models. To address both of these problems, a new concept of developing emulators for dynamic models is proposed. This concept is based on a prior that combines a simplified linear state space model of the temporal evolution of the dynamic model with Gaussian stochastic processes for the innovation terms as functions of model parameters and/or inputs. These innovation terms are intended to correct the error of the linear model at each output step. Conditioning this prior to the design data set is done by Kalman smoothing. This leads to an efficient emulator that, due to the consideration of our knowledge about dominant mechanisms built into the simulation model, can be expected to outperform purely statistical emulators at least in cases in which the design data set is small. The feasibility and potential difficulties of the proposed approach are demonstrated by the application to a simple hydrological model.

Distribution state estimation based on particle swarm and doubly loop mutant optimization (DLM-PSO)

This paper presents a novel algorithm based on particle swarm optimization (PSO) to estimate the states of electric distribution networks. In order to improve the performance, accuracy, convergence speed, and eliminate the stagnation effect of original PSO, a secondary PSO loop and mutation algorithm as well as stretching function is proposed. For accounting uncertainties of loads in distribution networks, pseudo-measurements is modeled as loads with the realistic errors. Simulation results on 6-bus radial and 34-bus IEEE test distribution networks show that the distribution state estimation based on proposed DLM-PSO presents lower estimation error and standard deviation in comparison with algorithms such as WLS, GA, HBMO, and original PSO.

Orientation robust transmit polarization diversity techniques

Polarisation diversity is a technique to improve the quality of mobile communications, but its reliability is suboptimal because it depends on the mobile channel and the antenna orientations at both ends of the mobile link. A method to optimise the reliability is established by minimising the dependency on antenna orientations. While the mobile base station can have fixed antenna orientation, the mobile terminal is typically a handheld device with random orientations. This means orientation invariance needs to be established at the receiver in the downlink, and at the transmitter in the uplink. This research presents separate solutions for both cases, and is based on the transmission of an elliptically polarised signal synthesised from the channel statistics. Complete receiver orientation invariance is achieved in the downlink. Effects of the transmitter orientation are minimised in the uplink.

Identification of abundant alkyl ether glycerophospholipids in the human lens by tandem mass spectrometry techniques

Previous studies have shown that the human lens contains glycerophospholipids with ether linkages. These lipids differ from conventional glycerophospholipids in that the sn-1 substituent is attached to the glycerol backbone via an 1-O-alkyl or an 1-O-alk-1'-enyl ether rather than an ester bond. The present investigation employed a combination of collision-induced dissociation (CID) and ozone-induced dissociation (OzID) to unambiguously distinguish such 1-O-alkyl and 1-O-alk-1'-enyl ethers. Using these methodologies the human lens was found to contain several abundant 1-O-alkyl glycerophos-phoethanolamines, including GPEtn(16:0e/9Z-18:1), GPEtn(11Z-18:1e/9Z-18:1), and GPEtn(18:0e/9Z-18:1), as well as a related series of unusual 1-O-alkyl glycerophosphoserines, including GPSer(16:0e/9Z-18:1), GPSer(11Z-18:1e/9Z-18:1), GPSer(18:0e/9Z-18:1) that to our knowledge have not previously been observed in human tissue. Isomeric 1-O-alk-1'-enyl ethers were absent or in low abundance. Examination of the double bond position within the phospholipids using OzID revealed that several positional isomers were present, including sites of unsaturation at the n-9, n-7, and even n-5 positions. Tandem CID/OzID experiments revealed a preference for double bonds in the n-7 position of 1-O-ether linked chains, while n-9 double bonds predominated in the ester-linked fatty acids [e.g., GPEtn(11Z-18:1e/9Z-18:1) and GPSer(11Z-18:1e/9Z-18:1)]. Different combinations of these double bond positional isomers within chains at the sn-1 and sn-2 positions point to a remarkable molecular diversity of ether-lipids within the human lens.

Critical analysis of spam prevention techniques

E-mail spam has remained a scourge and menacing nuisance for users, internet and network service operators and providers, in spite of the anti-spam techniques available; and spammers are relentlessly circumventing these anti-spam techniques embedded or installed in form of software products on both client and server sides of both fixed and mobile devices to their advantage. This continuous evasion degrades the capabilities of these anti-spam techniques as none of them provides a comprehensive reliable solution to the problem posed by spam and spammers. Major problem for instance arises when these anti-spam techniques misjudge or misclassify legitimate emails as spam (false positive); or fail to deliver or block spam on the SMTP server (false negative); and the spam passes-on to the receiver, and yet this server from where it originates does not notice or even have an auto alert service to indicate that the spam it was designed to prevent has slipped and moved on to the receiver’s SMTP server; and the receiver’s SMTP server still fail to stop the spam from reaching user’s device and with no auto alert mechanism to inform itself of this inability; thus causing a staggering cost in loss of time, effort and finance. This paper takes a comparative literature overview of some of these anti-spam techniques, especially the filtering technological endorsements designed to prevent spam, their merits and demerits to entrench their capability enhancements, as well as evaluative analytical recommendations that will be subject to further research.

Binary image steganographic techniques classification based on multi-class steganalysis

In this paper, we propose a new multi-class steganalysis for binary image. The proposed method can identify the type of steganographic technique used by examining on the given binary image. In addition, our proposed method is also capable of differentiating an image with hidden message from the one without hidden message. In order to do that, we will extract some features from the binary image. The feature extraction method used is a combination of the method extended from our previous work and some new methods proposed in this paper. Based on the extracted feature sets, we construct our multi-class steganalysis from the SVM classifier. We also present the empirical works to demonstrate that the proposed method can effectively identify five different types of steganography.