An augmented EDA with dynamic diversity control and local neighborhood search for coevolution of optimal negotiation strategies

In this paper, we present an estimation of distribution algorithm (EDA) augmented with enhanced dynamic diversity controlling and local improvement methods to solve competitive coevolution problems for agent-based automated negotiations. Since optimal negotiation strategies ensure that interacting agents negotiate optimally, finding such strategies—particularly, for the agents having incomplete information about their opponents—is an important and challenging issue to support agent-based automated negotiation systems. To address this issue, we consider the problem of finding optimal negotiation strategies for a bilateral negotiation between self-interested agents with incomplete information through an EDA-based coevolution mechanism. Due to the competitive nature of the agents, EDAs should be able to deal with competitive coevolution based on two asymmetric populations each consisting of self-interested agents. However, finding optimal negotiation solutions via coevolutionary learning using conventional EDAs is difficult because the EDAs suffer from premature convergence and their search capability deteriorates during coevolution. To solve these problems, even though we have previously devised the dynamic diversity controlling EDA (D²C-EDA), which is mainly characterized by a diversification and refinement (DR) procedure, D²C-EDA suffers from the population reinitialization problem that leads to a computational overhead. To reduce the computational overhead and to achieve further improvements in terms of solution accuracy, we have devised an improved D²C-EDA (ID²C-EDA) by adopting an enhanced DR procedure and a local neighborhood search (LNS) method. Favorable empirical results support the effectiveness of the proposed ID²C-EDA compared to conventional and the other proposed EDAs. Furthermore, ID²C-EDA finds solutions very close to the optimum.     

Superiority demonstration of variance-considered machines by comparing error rate with support vector machines

To improve the performance of classification algorithms, we proposed a new varianceconsidered machine (VCM) classification algorithm in a previous study. The study showed theoretically that VCMs have lower error probabilities than SVMs. The purpose of this paper is to experimentally demonstrate the superiority of VCMs. Therefore, we verified our proposal with several case experiments using data following a Gaussian distribution with different variances and prior probabilities. To estimate performance, the experiment for each case was executed 1000 times and the error rates were averaged for accuracy. The data of each experiment have different distances between means of data, and different ratios between training data and testing data. Thus, we proved that the error rate of VCMs is lower than the error rate of SVMs, although their performances were not similar in each case. Consequently, we expect that VCMs will be applied to a variety fields.     

Machine learning techniques for anti-money laundering (AML) solutions in suspicious transaction detection: a review

Money laundering has been affecting the global economy for many years. Large sums of money are laundered every year, posing a threat to the global economy and its security. Money laundering encompasses illegal activities that are used to make illegally acquired funds appear legal and legitimate. This paper aims to provide a comprehensive survey of machine learning algorithms and methods applied to detect suspicious transactions. In particular, solutions of anti-money laundering typologies, link analysis, behavioural modelling, risk scoring, anomaly detection, and geographic capability have been identified and analysed. Key steps of data preparation, data transformation, and data analytics techniques have been discussed; existing machine learning algorithms and methods described in the literature have been categorised, summarised, and compared. Finally, what techniques were lacking or under-addressed in the existing research has been elaborated with the purpose of pinpointing future research directions.     

A bioreactor with an internal contactor for primary recovery of bovine serum albumin from yeast suspension

In this paper the internal contactor is a newly developed device for the primary recovery of protein from crude feedstock. Ion exchanges (DEAE-Streamline) are confined inside the internal contactor in a stirred tank. Interactions between the ion exchange in the internal contactor and protein (BSA) in yeast suspension have been studied. For better performance, two strategies are considered: to determine the ion conductivity of a simulated yeast suspension, and to select the optimum process time for adsorption. In this system, advantages of both batch adsorption and expanded bed adsorption were obtained. Furthermore, in denser cell concentration (50 g/L) where EBA cannot be operated, the primary recovery was carried out in 1–2 hr. The efficiency of yield is higher than 80% in this condition.     

SET/CMOS hybrid process and multiband filtering circuits

We have developed an integration technology for the single electron transistor (SET)/CMOS hybrid systems. SET and CMOS transistors can be optimized without any possible degradation due to mixing dissimilar devices by adopting just one extra mask step for the separate gate oxidation (SGOX). We have confirmed that discrete devices show ideal characteristics required for the SET/CMOS hybrid systems. An SET shows obvious Coulomb oscillations with a 200-mV period and CMOS transistors show high voltage gain. Based on the hybrid process, new hybrid circuits, called periodic multiband filters, are proposed and successfully implemented. The new filter is designed to perform a filtering operation according to the periodic multiple blocking bands of which a period is originated from the SET. Such a novel function was implemented efficiently with a few transistors by making full use of the periodic nature of SET characteristics.     

Analysis of the HVAC system’s sound quality using the design of experiments

Human hearing is very sensitive to sound, so a subjective index of sound quality is required. Each situation of sound evaluation is composed of Sound Quality (SQ) metrics. When substituting the level of one frequency band, we could not see the tendency of substitution at the whole frequency band during SQ evaluation. In this study, the Design of Experiments (DOE) is used to analyze noise from an automotive Heating, Ventilating, and Air Conditioning (HVAC) system. The frequency domain is divided into 12 equal parts, and each level of the domain is given an increase or decrease due to the change in frequency band based on the “loud” and “sharp” sound of the SQ analyzed. By using DOE, the number of tests is effectively reduced by the number of experiments, and the main result is a solution at each band. SQ in terms of the “loud” and “sharp” sound at each band, the change in band (increase or decrease in sound pressure) or no change in band will have the most effect on the identifiable characteristics of SQ. This will enable us to select the objective frequency band. Through the results obtained, the physical level changes in arbitrary frequency domain sensitivity can be determined.     

Monolithically integrated grating cavity tunable lasers

A novel grating cavity tunable laser is proposed and experimentally demonstrated. It is realized by monolithically integrating a semiconductor optical amplifier, a dispersive element, a phase-control section, and an etched diffraction grating in a single chip. Tuning operation is based on electrically controlled beam deflection provided by dispersive element and phase matching by phase-control section. With the electrical control, the wavelength tuning of 8.5 nm with a sidemode suppression ratio of about 35 dB has been successfully achieved.     

Ontology-driven hypothesis generation to explain anomalous patient responses to treatment

Within the medical domain there are clear expectations as to how a patient should respond to treatments administered. When these responses are not observed it can be challenging for clinicians to understand the anomalous responses. The work reported here describes a tool which can detect anomalous patient responses to treatment and further suggest hypotheses to explain the anomaly. In order to develop this tool, we have undertaken a study to determine how Intensive Care Unit (ICU) clinicians identify anomalous patient responses; we then asked further clinicians to provide potential explanations for such anomalies. The high level reasoning deployed by the clinicians has been captured and generalised to form the procedural component of the ontology-driven tool. An evaluation has shown that the tool successfully reproduced the clinician’s hypotheses in the majority of cases. Finally, the paper concludes by describing planned extensions to this work.