Search

Liver transplantation (LT) for small infants remains challenging because of the demands related to graft selection, surgical technique, and perioperative management. The aim of this study was to evaluate the short‐term and longterm outcomes of LT regarding vascular/biliary complications, renal function, growth, and patient/graft survival in infants ≤3 months compared with those of an age between >3 and 6 months at a single transplant center. A total of 64 infants ≤6 months underwent LT and were divided into 2 groups according to age at LT: those of age ≤3 months (range, 6‐118 days; XS group, n = 37) and those of age >3 to ≤6 months (range, 124‐179 days; S group, n = 27) between 1989 and 2014. Acute liver failure was the main indication for LT in the XS group (n = 31, 84%) versus S (n = 7, 26%). The overall incidence of hepatic artery thrombosis and portal vein thrombosis/stricture were 5.4% and 10.8% in the XS group and 7.4% and 11.1% in the S group, respectively (not significant). The overall incidence of biliary stricture and leakage were 5.4% and 2.7% in the XS group and 3.7% and 3.7% in the S group, respectively (not significant). There was no significant difference between the 2 groups in terms of renal function. No significant difference was found between the 2 groups for each year after LT in terms of height and weight z score. The 1‐, 5‐, and 10‐year patient survival rates were 70.3%, 70.3%, and 70.3% in the XS group compared with 92.6%, 88.9%, and 88.9% in the S group, respectively (not significant). In conclusion, LT for smaller infants has acceptable outcomes despite the challenges of surgical technique, including vascular reconstruction and graft preparation, and perioperative management.

The past five years have seen a rapid development of plans and test pilots aimed at introducing connected and autonomous vehicles (CAVs) in public transport systems around the world. Using a real-world scenario from the Leeds Metropolitan Area as a case study, we demonstrate an effective way to combine macro-level mobility simulations based on open data (i.e., geographic information system information and transit timetables) with evolutionary optimisation techniques to discover realistic optimised integration routes for CAVs. The macro-level mobility simulations are used to assess the quality (i.e., fitness) of a potential CAV route by quantifying geographic accessibility improvements using an extended version of Dijkstra's algorithm on an abstract multi-modal transport network.

In recent years, deep learning has rapidly become a method of choice for segmentation of medical images. Deep neural architectures such as UNet and FPN have achieved high performances on many medical datasets. However, medical image analysis algorithms are required to be reliable, robust, and accurate for clinical applications which can be difficult to achieve for some single deep learning methods. In this study, we introduce an ensemble of classifiers for semantic segmentation of medical images. The ensemble of classifiers here is a set of various deep learning-based classifiers, aiming to achieve better performance than using a single classifier. We propose a weighted ensemble method in which the weighted sum of segmentation outputs by classifiers is used to choose the final segmentation decision. We use a swarm intelligence algorithm namely Comprehensive Learning Particle Swarm Optimization to optimize the combining weights. Dice coefficient, a popular performance metric for image segmentation, is used as the fitness criteria. Experiments conducted on some medical datasets of the CAMUS competition on cardiographic image segmentation show that our method achieves better results than both the constituent segmentation models and the reported model of the CAMUS competition.

In this study, we introduce an ensemble selection method for deep ensemble systems called VEGAS. The deep ensemble models include multiple layers of the ensemble of classifiers (EoC). At each layer, we train the EoC and generates training data for the next layer by concatenating the predictions for training observations and the original training data. The predictions of the classifiers in the last layer are combined by a combining method to obtain the final collaborated prediction. We further improve the prediction accuracy of a deep ensemble model by searching for its optimal configuration, i.e., the optimal set of classifiers in each layer. The optimal configuration is obtained using the Variable-Length Genetic Algorithm (VLGA) to maximize the prediction accuracy of the deep ensemble model on the validation set. We developed three operators of VLGA: roulette wheel selection for breeding, a chunk-based crossover based on the number of classifiers to generate new offsprings, and multiple random points-based mutation on each offspring. The experiments on 20 datasets show that VEGAS outperforms selected benchmark algorithms, including two well-known ensemble methods (Random Forest and XgBoost) and three deep learning methods (Multiple Layer Perceptron, gcForest, and MULES).

Text classification is one of the most important tasks in the field of Natural Language Processing. There are many approaches that focus on two main aspects: generating an effective representation; and selecting and refining algorithms to build the classification model. Traditional machine learning methods represent documents in vector space using features such as term frequencies, which have limitations in handling the order and semantics of words. Meanwhile, although achieving many successes, deep learning classifiers require substantial resources in terms of labelled data and computational complexity. In this work, a weighted ensemble of classifiers (WEC) is introduced to address the text classification problem. Instead of using majority vote as the combining method, we propose to associate each classifier’s prediction with a different weight when combining classifiers. The optimal weights are obtained by minimising a loss function on the training data with the Particle Swarm Optimisation algorithm. We conducted experiments on 5 popular datasets and report classification performance of algorithms with classification accuracy and macro F1 score. WEC was run with several different combinations of traditional machine learning and deep learning classifiers to show its flexibility and robustness. Experimental results confirm the advantage of WEC, especially on smaller datasets.

Deep Neural Networks have achieved many successes when applying to visual, text, and speech information in various domains. The crucial reasons behind these successes are the multi-layer architecture and the in-model feature transformation of deep learning models. These design principles have inspired other sub-fields of machine learning including ensemble learning. In recent years, there are some deep homogenous ensemble models introduced with a large number of classifiers in each layer. These models, thus, require a costly computational classification. Moreover, the existing deep ensemble models use all classifiers including unnecessary ones which can reduce the predictive accuracy of the ensemble. In this study, we propose a multi-layer ensemble learning framework called MUlti-Layer heterogeneous Ensemble System (MULES) to solve the classification problem. The proposed system works with a small number of heterogeneous classifiers to obtain ensemble diversity, therefore being efficiency in resource usage. We also propose an Evolutionary Algorithm-based selection method to select the subset of suitable classifiers and features at each layer to enhance the predictive performance of MULES. The selection method uses NSGA-II algorithm to optimize two objectives concerning classification accuracy and ensemble diversity. Experiments on 33 datasets confirm that MULES is better than a number of well-known benchmark algorithms.

Combining classifiers in an ensemble is beneficial in achieving better prediction than using a single classifier. Furthermore, each classifier can be associated with a weight in the aggregation to boost the performance of the ensemble system. In this work, we propose a novel dynamic weighted ensemble method. Based on the observation that each classifier provides a different level of confidence in its prediction, we propose to encode the level of confidence of a classifier by associating with each classifier a credibility threshold, computed from the entire training set by minimizing the entropy loss function with the mini-batch gradient descent method. On each test sample, we measure the confidence of each classifier’s output and then compare it to the credibility threshold to determine whether a classifier should be attended in the aggregation. If the condition is satisfied, the confidence level and credibility threshold are used to compute the weight of contribution of the classifier in the aggregation. By this way, we are not only considering the presence but also the contribution of each classifier based on the confidence in its prediction on each test sample. The experiments conducted on a number of datasets show that the proposed method is better than some benchmark algorithms including a non-weighted ensemble method, two dynamic ensemble selection methods, and two Boosting methods.

The field of object detection has witnessed great strides in recent years. With the wave of deep neural networks (DNN), many breakthroughs have achieved for the problems of object detection which previously were thought to be difficult. However, there exists a limitation with DNN-based approaches as some architectures are only suitable for particular types of object. Thus it would be desirable to combine the strengths of different methods to handle objects in different contexts. In this study, we propose an ensemble of object detectors in which individual detectors are adaptively combine for the collaborated decision. The combination is conducted on the outputs of detectors including the predicted label and location for each object. We proposed a detector selection method to select the suitable detectors and a weighted-based combining method to combine the predicted locations of selected detectors. The parameters of these methods are optimized by using Particle Swarm Optimization in order to maximize mean Average Precision (mAP) metric. Experiments conducted on VOC2007 dataset with six object detectors show that our ensemble method is better than each single detector.

In ensemble systems, the predictions of base classifiers are aggregated by a combining algorithm (meta-classifier) to achieve better classification accuracy than using a single classifier. Experiments show that the performance of ensembles significantly depends on the choice of meta-classifier. Normally, the classifier selection method applied to an ensemble usually removes all the predictions of a classifier if this classifier is not selected in the final ensemble. Here we present an idea to only remove a subset of each classifier's prediction thereby introducing a simultaneous meta-data and meta-classifier selection method for ensemble systems. Our approach uses Cross Validation on the training set to generate meta-data as the predictions of base classifiers. We then use Ant Colony Optimization to search for the optimal subset of meta-data and meta-classifier for the data. By considering each column of meta-data, we construct the configuration including a subset of these columns and a meta-classifier. Specifically, the columns are selected according to their corresponding pheromones, and the meta-classifier is chosen at random. The classification accuracy of each configuration is computed based on Cross Validation on meta-data. Experiments on UCI datasets show the advantage of proposed method compared to several classifier and feature selection methods for ensemble systems.

Benchmark sets and landscape features are used to test algorithms and to train models to perform algorithm selection or configuration. These approaches are based on the assumption that algorithms have similar performances on problems with similar feature sets. In this paper, we test different configurations of differential evolution (DE) against the BBOB set. We then use the landscape features of those problems and a case base reasoning approach for DE configuration selection. We show that, although this method obtains good results for BBOB problems, it fails to select the best configurations when facing a new set of optimisation problems with a distinct array of landscape features. This demonstrates the limitations of the BBOB set for algorithm selection. Moreover, by examination of the relationship between features and algorithm performance, we show that there is no correlation between the feature space and the performance space. We conclude by identifying some important open questions raised by this work.

We present a new variant of the Vehicle Routing Problem based on a real industrial scenario. This VRP is dynamic and heavily constrained and uses time-windows, a heterogeneous vehicle fleet and multiple types of job. A constructive solver is developed and tested using dynamic simulation of real-world data from a leading Scottish haulier. Our experiments establish the efficiency and reliability of the method for this problem. Additionally, a methodology for evaluating policy changes through simulation is presented, showing that our technique supports operations and management. We establish that fleet size can be reduced or more jobs handled by the company.

Designing an ensemble of classifiers is one of the popular research topics in machine learning since it can give better results than using each constituent member. Furthermore, the performance of ensemble can be improved using selection or adaptation. In the former, the optimal set of base classifiers, meta-classifier, original features, or meta-data is selected to obtain a better ensemble than using the entire classifiers and features. In the latter, the base classifiers or combining algorithms working on the outputs of the base classifiers are made to adapt to a particular problem. The adaptation here means that the parameters of these algorithms are trained to be optimal for each problem. In this study, we propose a novel evolving combining algorithm using the adaptation approach for the ensemble systems. Instead of using numerical value when computing the representation for each class, we propose to use the interval-based representation for the class. The optimal value of the representation is found through Particle Swarm Optimization. During classification, a test instance is assigned to the class with the interval-based representation that is closest to the base classifiers’ prediction. Experiments conducted on a number of popular dataset confirmed that the proposed method is better than the well-known ensemble systems using Decision Template and Sum Rule as combiner, L2-loss Linear Support Vector Machine, Multiple Layer Neural Network, and the ensemble selection methods based on GA-Meta-data, META-DES, and ACO.

In this paper, we aim to develop an effective combining algorithm for ensemble learning systems. The Decision Template method, one of the most popular combining algorithms for ensemble systems, does not perform well when working on certain datasets like those having imbalanced data. Moreover, point estimation by computing the average value on the outputs of base classifiers in the Decision Template method is sometimes not a good representation, especially for skewed datasets. Here we propose to search for an optimal decision template in the combining algorithm for a heterogeneous ensemble. To do this, we first generate the base classifier by training the pre-selected learning algorithms on the given training set. The meta-data of the training set is then generated via cross validation. Using the Artificial Bee Colony algorithm, we search for the optimal template that minimizes the empirical 0–1 loss function on the training set. The class label is assigned to the unlabeled sample based on the maximum of the similarity between the optimal decision template and the sample’s meta-data. Experiments conducted on the UCI datasets demonstrated the superiority of the proposed method over several benchmark algorithms.

Ensemble selection is one of the most studied topics in ensemble learning because a selected subset of base classifiers may perform better than the whole ensemble system. In recent years, a great many ensemble selection methods have been introduced. However, many of these lack flexibility: either a fixed subset of classifiers is pre-selected for all test samples (static approach), or the selection of classifiers depends upon the performance of techniques that define the region of competence (dynamic approach). In this paper, we propose an ensemble selection method that takes into account each base classifier's confidence during classification and the overall credibility of the base classifier in the ensemble. In other words, a base classifier is selected to predict for a test sample if the confidence in its prediction is higher than its credibility threshold. The credibility thresholds of the base classifiers are found by minimizing the empirical 0–1 loss on the entire training observations. In this way, our approach integrates both the static and dynamic aspects of ensemble selection. Experiments on 62 datasets demonstrate that the proposed method achieves much better performance in comparison to some ensemble methods.

By incorporating domain knowledge, simple greedy procedures can be defined to generate reasonably good solutions to many optimisation problems. However, such solutions are unlikely to be optimal and their quality often depends on the way the decision variables are input to the greedy method. Indirect optimisation uses meta-heuristics to optimise the input of the greedy decoders. As the performance and the runtime differ across greedy methods and meta-heuristics, deciding how to split the computational effort between the two sides of the optimisation is not trivial and can significantly impact the search. In this paper, an artificial scheduling problem is presented along with five greedy procedures, using varying levels of domain information. A methodology to compare different indirect optimisation strategies is presented using a simple Hill Climber, a Genetic Algorithm and a population-based Local Search. By assessing all combinations of meta-heuristics and greedy procedures on a range of problem instances with different properties, experiments show that encapsulating problem knowledge within greedy decoders may not always prove successful and that simpler methods can lead to comparable results as advanced ones when combined with meta-heuristics that are adapted to the problem. However, the use of efficient greedy procedures reduces the relative difference between meta-heuristics.

The service chain planning process is a critical component in the operations of companies in the service industry, such as logistics, telecoms or utilities. This process involves looking ahead over various timescales to ensure that available capacity matches the required demand whilst maximizing revenues and minimizing costs. This problem is particularly complex for companies with large, multi-skilled workforces as matching these resources to the required demand can be done in a vast number of combinations. The vastness of the problem space combined with the criticality to the business is leading to an increasing move towards automation of the process in recent years. In this paper we focus on the tactical plan where planning is occurring daily for the coming weeks, matching the available capacity to demand, using capacity levers to flex capacity to keep backlogs within target levels whilst maintaining target levels for provision of new revenues. First we describe the tactical planning problem before defining a bi-level model to search for optimal solutions to it. We show, by comparing the model results to actual planners on real world examples, that the bi-level model produces good results that replicate the planners' process whilst keeping the backlogs closer to target levels, thus providing a strong case for its use in the automation of the tactical planning process.

The Artificial Reaction Network (ARN) is an artificial chemistry inspired by Cell Signalling Networks. Its purpose is to represent chemical circuitry and to explore the computational properties responsible for generating emergent high-level behaviour. In this paper, the design and application of ARN-based cell-like agents termed "Cytobots" are explored. Such agents provide a facility to explore the dynamics and emergent properties of multicellular systems. The Cytobot ARN is constructed by combining functional motifs found in real biochemical networks. By instantiating this ARN, multiple Cytobots are created, each of which is capable of recognising environmental patterns, stigmergic communication with others and controlling its own trajectory. Applications in biological simulation and robotics are investigated by first applying the agents to model the life-cycle phases of the cellular slime mould D. discoideum and then to simulate an oil-spill clean-up operation. The results demonstrate that an ARN-based approach provides a powerful tool for modelling multi-agent biological systems and also has application in swarm robotics.

The National Emerging Infectious Diseases Laboratories (NEIDL), Boston University, is a globally unique biocontainment research facility housing biosafety level 2 (BSL-2), BSL-3, and BSL-4 laboratories. Located in the BioSquare area at the University's Medical Campus, it is part of a national network of secure facilities constructed to study infectious diseases of major public health concern. The NEIDL allows for basic, translational, and clinical phases of research to be carried out in a single facility with the overall goal of accelerating understanding, treatment, and prevention of infectious diseases. The NEIDL will also act as a center of excellence providing training and education in all aspects of biocontainment research. Within every detail of NEIDL operations is a primary emphasis on safety and security. The ultramodern NEIDL has required a new approach to communications technology solutions in order to ensure safety and security and meet the needs of investigators working in this complex building. This article discusses the implementation of secure wireless networks and private cloud computing to promote operational efficiency, biosecurity, and biosafety with additional energy-saving advantages. The utilization of a dedicated data center, virtualized servers, virtualized desktop integration, multichannel secure wireless networks, and a NEIDL-dedicated Voice over Internet Protocol (VoIP) network are all discussed.

Fitness modeling has received growing interest from the evolutionary computation community in recent years. With a fitness model, one can improve evolutionary algorithm efficiency by directly sampling new solutions, developing hybrid guided evolutionary operators or using the model as a surrogate for an expensive fitness function. This paper addresses several issues on fitness modeling of discrete functions, particularly how modeling quality and efficiency can be improved. We define the Markov network fitness model in terms of Walsh functions. We explore the relationship between the Markov network fitness model and fitness in a number of discrete problems, showing how the parameters of the fitness model can identify qualitative features of the fitness function. We define the fitness prediction correlation, a metric to measure fitness modeling capability of local and global fitness models. We use this metric to investigate the effects of population size and selection on the tradeoff between model quality and complexity for the Markov network fitness model.

The Artificial Reaction Network (ARN) is a Cell Signalling Network inspired connectionist representation belonging to the branch of A-Life known as Artificial Chemistry. Its purpose is to represent chemical circuitry and to explore computational properties responsible for generating emergent high-level behaviour associated with cells. In this paper, the computational mechanisms involved in pattern recognition and spatio-temporal pattern generation are examined in robotic control tasks. The results show that the ARN has application in limbed robotic control and computational functionality in common with Artificial Neural Networks. Like spiking neural models, the ARN can combine pattern recognition and complex temporal control functionality in a single network, however it offers increased flexibility. Furthermore, the results illustrate parallels between emergent neural and cell intelligence.

In this paper we present a method to predict service levels in utility companies, giving them advanced visibility of expected service outcomes and helping them to ensure adherence to service level agreements made to their clients. Service level adherence is one of the key targets during the service chain planning process in service industries, such as telecoms or utility companies. These specify a time limit for successful completion of a certain percentage of tasks on that service level agreement. With the increasing use of automation within the planning process, the requirement for a method to evaluate the current plan decisions effects on service level outcomes has surfaced.

The Multi-mode Resource Constrained Project Scheduling Problem (MRCPSP) has been of research interest for over two decades. The problem is composed of two interacting sub problems: mode assignment and activity scheduling. These problems cannot be solved in isolation because of the interaction that exists between them. Many evolutionary algorithms have been applied to this problem most commonly the Genetic Algorithm (GA). It has been common practice to improve the performance of the GA with some local search techniques. The Bi-population Genetic Algorithm (BPGA) is one of the most competitive GAs for solving the MRCPSP. In this paper, we improve the BPGA by hybridising it with an Estimation of Distribution Algorithm that focuses on improving how modes are generated. We also suggest improvement to the existing experimental methodology.

Critical for successful operations of service industries, such as telecoms, utility companies and logistic companies, is the service chain planning process. This involves optimizing resources against expected demand to maximize the utilization and minimize the wastage, which in turn maximizes revenue whilst minimizing the cost. This is increasingly involving the automation of the planning process. However, due to unforeseen factors, the calculated optimal allocation of resources to complete tasks often does not match up with what is actually occurring on the day. This factor highlights a requirement for a method of predicting accurately the number of tasks that will be completed given a known amount of resources and demand in order to produce a more accurate plan.

Objectives: To construct new prostate cancer staging lookup tables based on a dataset collated by the British Association of Urological Surgeons (BAUS) and to validate them and compare their predictive power with Partin tables. Patients and methods: Complete data on 1701 patients was collated between 1999 and 2008 across 57 UK centres. Lookup tables were created for prediction of pathological stage (PS) using PSA level, biopsy Gleason score (GS) and clinical stage, replicating Partin's original approach. Tables were generated using logistic regression (LR) and bootstrap resampling methods and were internally validated and externally validated using concordance indices (CI) and area under the receiver operating characteristic curve (AUROC) respectively. Results: The CI and AUROC analyses indicate that Partin tables performed poorly on UK data in comparison with US data. The UK prostate cancer tables performed better than Partin tables but the predictive power of all models was relatively poor. Conclusion: The study shows that the predictive power of Partin tables is reduced when applied to the UK population. Models generated using LR methodology have fundamental limitations, and we suggest alternative modelling methods such as Bayesian networks.

Objectives: Prediction of prostate cancer pathological stage is an essential step in a patient's pathway. It determines the treatment that will be applied further. In current practice, urologists use the pathological stage predictions provided in Partin tables to support their decisions. However, Partin tables are based on logistic regression (LR) and built from US data. Our objective is to investigate a range of both predictive methods and of predictive variables for pathological stage prediction and assess them with respect to their predictive quality based on UK data. Methods and material: The latest version of Partin tables was applied to a large scale British dataset in order to measure their performances by mean of concordance index (c-index). The data was collected by the British Association of Urological Surgeons (BAUS) and gathered records from over 1700 patients treated with prostatectomy in 57 centers across UK. The original methodology was replicated using the BAUS dataset and evaluated using concordance index. In addition, a selection of classifiers, including, among others, LR, artificial neural networks and Bayesian networks (BNs) was applied to the same data and compared with each other using the area under the ROC curve (AUC). Subsets of the data were created in order to observe how classifiers perform with the inclusion of extra variables. Finally a local dataset prepared by the Aberdeen Royal Infirmary was used to study the effect on predictive performance of using different variables. Results: Partin tables have low predictive quality (c-index=0.602) when applied on UK data for comparison on patients with organ confined and extra prostatic extension conditions, patients at the two most frequently observed pathological stages. The use of replicate lookup tables built from British data shows an improvement in the classification, but the overall predictive quality remains low (c-index=0.610). Comparing a range of classifiers shows that BNs generally outperform other methods. Using the four variables from Partin tables, naive Bayes is the best classifier for the prediction of each class label (AUC=0.662 for OC). When two additional variables are added, the results of LR (0.675), artificial neural networks (0.656) and BN methods (0.679) are overall improved. BNs show higher AUCs than the other methods when the number of variables raises Conclusion: The predictive quality of Partin tables can be described as low to moderate on UK data. This means that following the predictions generated by Partin tables, many patients would received an inappropriate treatment, generally associated with a deterioration of their quality of life. In addition to demographic differences between UK and the original US population, the methodology and in particular LR present limitations. BN represents a promising alternative to LR from which prostate cancer staging can benefit. Heuristic search for structure learning and the inclusion of more variables are elements that further improve BN models quality.

Transient ischaemic attack (TIA) and stroke are clinical syndromes characterised by acute neurological deficits with vascular causes. People experiencing TIA or a first stroke are at significant risk of subsequent stroke. Risk factors have been identified and include factors associated with lifestyle such as tobacco use, diet, obesity, alcohol consumption, physical activity and stress. Targeted therapeutic interventions have the potential to reduce the burden of stroke substantially. The aim of this article is to provide an overview of the evidence relating to lifestyle risk factors for stroke. Health promotion theories and intervention techniques that nurses can use to address lifestyle behaviour change following stroke will also be discussed. Authors

The need to assess internet performance from the user’s perspective grows, as does the interest in deployment of Large-Scale Measurement Platforms (LMAPs). The potential of these platforms as a real-time network diagnostic tool is limited by the volume, velocity and variety of the data they generated. Fusing this data from multiple sources and generating a single piece of coherent information about the state of the network would increase the efficiency of network monitoring. The current practice of visually analysing LMAPs’ data stream would certainly benefit from having automatically generated notifications in a timely manner alerting human controllers to the network’s conditions of interest. This paper proposed a data fusion framework for LMAPs that makes use of mathematical distribution based sensors to generate probabilistic sensor outputs which are fused using a Dempster- Shafer Theory.

Structural learning of Bayesian networks is a very expensive task even when sacrifying the optimality of the result. Because of that, there are some proposals aimed at obtaining relative-quality solutions in short times. One of them, namely Chain-ACO, searches an ordering among all variables with Ant Colony Optimization and a chain-structured surrogate model, and then uses this ordering to build a Bayesian network by means of the well-known K2 algorithm.This work is based on Chain-ACO. We evaluate the impact of using a tree-structured surrogate model instead of a chain to evaluate orderings during the search. Moreover, we propose a variation of the way K2 builds the network, which consists of allowing some changes in the positions of the variables whenever they such changes not produce a cycle. This modification of the process may improve the score of the final network, without (almost) any additional cost.

Genetic algorithms (GAs) are a heuristic search and optimisation technique inspired by natural evolution. They have been successfully applied to a wide range of real-world problems of significant complexity. This paper is intended as an introduction to GAs aimed at immunologists and mathematicians interested in immunology. We describe how to construct a GA and the main strands of GA theory before speculatively identifying possible applications of GAs to the study of immunology. An illustrative example of using a GA for a medical optimal control problem is provided. The paper also includes a brief account of the related area of artificial immune systems.

Background Little is known about the pattern of injury in short track speed skating. Purpose To investigate the incidence and characteristics of injuries in short track speed skating. Study Design Retrospective study. Methods Ninety-five of 150 elite-level skaters (63.3%) were surveyed to collect information on training and competition load as well as on injuries sustained during the 1999—2000 competitive season. Injuries were characterized in terms of anatomic location, type of injury, time loss from training and competition, and circumstance of injury (acute onset during competition, on-ice practice, off-ice training, or insidious onset). Results Sixty-one of the 95 skaters (64.2%) reported sustaining at least one injury. The knee, ankle, spine, leg, and groin were the most commonly reported sites of injury. Skaters were also asked to list previous on-ice injuries. The two most common injuries occurring on-ice before the 1999—2000 season were lacerations from the knee down (11.1%) and ankle fractures (10.2%). Conclusion The results of this study suggest that there is a high incidence of injury in competitive short track speed skating.

Despite multiple publications on potential therapies for neuromuscular diseases (NMD) in cell and animal models only a handful reach clinical trials. The ability to prioritise drug development according to objective criteria is particularly critical in rare diseases with large unmet needs and a limited numbers of patients who can be enrolled into clinical trials. TREAT-NMD Advisory Committee for Therapeutics (TACT) was established to provide independent and objective guidance on the preclinical and development pathway of potential therapies (whether novel or repurposed) for NMD. We present our experience in the establishment and operation of the TACT. TACT provides a unique resource of recognized experts from multiple disciplines. The goal of each TACT review is to help the sponsor to position the candidate compound along a realistic and well-informed plan to clinical trials, and eventual registration. The reviews and subsequent recommendations are focused on generating meaningful and rigorous data that can enable clear go/no-go decisions and facilitate longer term funding or partnering opportunities. The review process thereby acts to comment on viability, de-risking the process of proceeding on a development programme. To date TACT has held 10 review meeting and reviewed 29 program applications in several rare neuromuscular diseases: Of the 29 programs reviewed, 19 were from industry and 10 were from academia; 15 were for novel compounds and 14 were for repurposed drugs; 16 were small molecules and 13 were biologics; 14 were preclinical stage applications and 15 were clinical stage applications. 3 had received Orphan drug designation from European Medicines Agency and 3 from Food and Drug Administration. A number of recurrent themes emerged over the course of the reviews and we found that applicants frequently require advice and education on issues concerned with preclinical standard operating procedures, interactions with regulatory agencies, formulation, repurposing, clinical trial design, manufacturing and ethics. Over the 5 years since its establishment TACT has amassed a body of experience that can be extrapolated to other groups of rare diseases to improve the community’s chances of successfully bringing new rare disease drugs to registration and ultimately to market.

Background The outcome of treatment for advanced hepatoblastoma has recently improved after the introduction of preoperative or pre- and postoperative cisplatin-containing chemotherapy combined with complete surgical excision. The role of liver transplantation in a population of patients who have received this regimen has not been clearly defined. Methods Orthotopic liver transplantation (OLT) was performed in 13 children, aged 5 months to 11 years (median 27 months), who were assessed with unresectable hepatoblastoma, and whose pretreatment extent-of-disease was based on radiologic findings of group III (n=11) and group IV (n=2). One child with a multifocal tumor showed pulmonary metastases at presentation, but, according to radiologic studies, the deposits resolved with chemotherapy before liver transplantation. One other child showed exophytic extension of the primary tumor infiltrating the porta hepatis and body of the pancreas. All 13 patients received preoperative chemotherapy to reduce the size of the primary tumor(s) and to treat metastatic spread. Results Twelve children underwent elective OLT; all are alive and show normal graft function at a mean follow-up of 33 months (range 1-108). One child shows evidence of recurrent disease in the form of pulmonary metastases. One child underwent emergency OLT for acute liver failure after (incomplete) extended right hepatectomy and died from respiratory failure, with no evidence of recurrent tumor 3 weeks posttransplant. Conclusions Liver transplantation is an effective treatment for unresectable unifocal or multifocal hepatoblastoma confined to the liver. A multidisciplinary approach to the management of hepatoblastoma, with thoughtful collaboration between pediatric oncologists, hepatologists, and liver surgeons, is essential.

Inflammatory pseudotumor-like follicular dendritic cell (FDC) tumor is an uncommon, Epstein-Barr virus-associated neoplasm of the liver or spleen, characterized by spindly tumor cells dispersed in a background of small lymphocytes and plasma cells. We report 6 diagnostically challenging cases in which the neoplastic component is further overshadowed by granulomas or eosinophils. The patients included 2 men and 4 women with a median age of 45.5 years, and 1 of them showed concurrent involvement of the liver and spleen. The presence of extensive coalescent epithelioid granulomas in 3 splenic tumors and 3 liver tumors raised the possibilities of an infective process or sarcoidosis. In another liver tumor, the massive infiltrate of eosinophils, accompanied by geographic eosinophilic abscesses, suggested parasitic infestation or so-called eosinophilic granuloma of the liver. However, scrutiny of the tissue between the granulomas or among the eosinophils revealed scattered atypical spindly cells with indistinct cell borders, large vesicular nuclei, and distinct nucleoli. The atypical cells were positive for FDC markers on immunostaining (CD21, CD35) and Epstein-Barr virus on in situ hybridization. Thus, a diagnosis of inflammatory pseudotumor-like FDC tumor could be confirmed. Awareness of the granulomatous and eosinophil-rich variants of this tumor type will facilitate the correct diagnosis to be made.

It is controversial whether prolonged antibiotic treatment is effective for patients in whom symptoms persist after the recommended antibiotic treatment for acute Lyme disease.We conducted two randomized trials: one in 78 patients who were seropositive for IgG antibodies to Borrelia burgdorferi at the time of enrollment and the other in 51 patients who were seronegative. The patients received either intravenous ceftriaxone, 2 g daily for 30 days, followed by oral doxycycline, 200 mg daily for 60 days, or matching intravenous and oral placebos. Each patient had well-documented, previously treated Lyme disease but had persistent musculoskeletal pain, neurocognitive symptoms, or dysesthesia, often associated with fatigue. The primary outcome measures were improvement on the physical- and mental-health-component summary scales of the Medical Outcomes Study 36-item Short-Form General Health Survey (SF-36)--a scale measuring the health-related quality of life--on day 180 of the study.After a planned interim analysis, the data and safety monitoring board recommended that the studies be discontinued because data from the first 107 patients indicated that it was highly unlikely that a significant difference in treatment efficacy between the groups would be observed with the planned full enrollment of 260 patients. Base-line assessments documented severe impairment in the patients' health-related quality of life. In intention-to-treat analyses, there were no significant differences in the outcomes with prolonged antibiotic treatment as compared with placebo. Among the seropositive patients who were treated with antibiotics, there was improvement in the score on the physical-component summary scale of the SF-36, the mental-component summary scale, or both in 37 percent, no change in 29 percent, and worsening in 34 percent; among seropositive patients receiving placebo, there was improvement in 40 percent, no change in 26 percent, and worsening in 34 percent (P=0.96 for the comparison between treatment groups). The results were similar for the seronegative patients.There is considerable impairment of health-related quality of life among patients with persistent symptoms despite previous antibiotic treatment for acute Lyme disease. However, in these two trials, treatment with intravenous and oral antibiotics for 90 days did not improve symptoms more than placebo.

Cefprozil is a beta-lactamase-stable oral cephalosporin with an antimicrobial spectrum that includes gram-positive and gram-negative pathogens commonly associated with acute bacterial sinusitis, one of the most common upper respiratory tract infections among adults. We conducted a multicenter, open-label study to compare the efficacy and safety of cefprozil and amoxicillin/clavulanate in the treatment of adults with severe acute bacterial sinusitis diagnosed by clinical and radiographic criteria. A total of 278 patients entered the study, 140 (59 males, 81 females) in the cefprozil group and 138 (69 males, 69 females) in the amoxicillin/clavulanate group. Patients were randomized to 10 days of treatment with either cefprozil 500 mg BID or amoxicillin/clavulanate 500 mg/125 mg TID. Clinical severity was assessed at study entry, and patients were stratified based on symptom grade. Efficacy was evaluated using a 10-point questionnaire administered during, at the end of, and 2 weeks after completing therapy. At the end of treatment, 84.5% (71/84) of patients with severe sinusitis treated with cefprozil had a satisfactory clinical response, which was not significantly different from the 89.9% (80/89) of patients in the amoxicillin/clavulanate group who had a satisfactory clinical response. Two weeks after completing treatment, 80.8% (63/78) of cefprozil-treated patients and 81.0% (64/79) of amoxicillin/clavulanate-treated patients with severe sinusitis had a satisfactory response. Relapse was more common among amoxicillin/clavulanate patients (6/70; 8.6%) than among cefprozil patients (2/65; 3.1%), but the difference was not statistically significant. Significantly more amoxicillin/clavulanate-treated patients experienced adverse events compared with cefprozil-treated patients (P < 0.001), including diarrhea (P < 0.001), nausea (P < 0.042), and rash (P < 0.035). Three times as many amoxicillin/clavulanate patients discontinued treatment because of adverse events. Cefprozil demonstrated comparable clinical efficacy to amoxicillin/clavulanate in the treatment of adults with severe sinusitis; however, cefprozil was associated with a significantly lower incidence of diarrhea, nausea, and rash.

A JAVA application, The Oncologists Workbench, which allows oncologists to estimate the influence of new cancer treatment schedules is being developed. The requirement for a rational approach to the design of chemotherapeutic regimens is well established [1]. Our prototype allows oncologists using the World Wide Web (WWW) to graphically construct treatment regimens while considering various toxic side effects. A simulation engine makes predictions of tumour growth based on previous clinical knowledge of response to treatment. The oncologist can then examine the predicted tumour response information with a specially constructed interactive viewer. These interlinked tools allow oncologists to develop and predict the effectiveness of novel chemotherapeutic regimens. This work is part of an ongoing collaboration between oncologists, mathematicians and computer scientists to provide tools for improving cancer chemotherapy.

The Artificial Reaction Network (ARN) is an Artificial Chemistry representation inspired by cell signaling networks. The ARN has previously been applied to the simulation of the chemotaxis pathway of Escherichia coli and to the control of limbed robots. In this paper we discuss the design of an ARN control system composed of a combination of network motifs found in actual biochemical networks. Using this control system we create multiple cell-like autonomous agents capable of coordinating all aspects of their behavior, recognizing environmental patterns and communicating with other agent's stigmergically. The agents are applied to simulate two phases of the life cycle of Dictyostelium discoideum: vegetative and aggregation phase including the transition. The results of the simulation show that the ARN is well suited for construction of biochemical regulatory networks. Furthermore, it is a powerful tool for modeling multi agent systems such as a population of amoebae or bacterial colony.

The Artificial Reaction Network (ARN) is a cell signaling network inspired representation belonging to the branch of A-Life known as Artificial Chemistry. It has properties in common with both AI and Systems Biology techniques including Artificial Neural Networks, Petri Nets, Random Boolean Networks and S-Systems. The ARN has been previously applied to control of limbed robots and simulation of biological signaling pathways. In this paper, multiple instances of independent distributed ARN controlled agents function to find the global minima within a set of simulated environments characterized by benchmark problems. The search behavior results from the internal ARN network, but is enhanced by collective activities and stigmergic interaction of the agents. The results show that the agents are able to find best fitness solutions in all problems, and compare well with results of cell inspired optimization algorithms. Such a system may have practical application in distributed or swarm robotics.

Agent-based simulation is increasingly used in industry to model systems of interest allowing the evaluation of alternative scenarios. By this means, business managers can estimate the consequences of policy changes at low cost before implementing them in the business. However, in order to apply such models with confidence, it is necessary to validate them continuously against changing business patterns. Typically, models contain key parameters which significantly affect the overall behaviour of the system. The process of selecting such parameters is an inverse problem known as ‘tuning’ In this chapter, we describe the application of computational intelligence to tune the parameters of a workforce dynamics simulator. We show that the best algorithm achieves reduced tuning times as well as more accurate field workforce simulations. Since implementation, this algorithm has facilitated the use of simulation to assess the effect of changes in different business scenarios and transformation initiatives.

Bayesian network (BN) structure learning is an NP hard problem. Search and score algorithms are one of the main approaches proposed for learning BN structure from data. Previous research has shown that the relative performances of such algorithms are problem dependent and that fitness landscape analysis can be used to characterize the difficulty of the search for different scoring functions. In this paper, we construct a classifier based on fitness landscape analysis and receiver operating characteristic curves. The classifier labels search landscapes with the most suitable scoring function. We train the classifier on a number of standard benchmark functions. The classifier forms the basis for a selective hyperheuristic algorithm. This uses an initial landscape analysis stage to select a scoring function using the classifier. The hyperheuristic algorithm is tested on a distribution of unseen problems based on mutations of the standard benchmarks. Our results establish that the hyperheuristic performs better than a uniformly random scoring function selection approach that omit the landscape analysis stage. Therefore the effects on performance of problem-dependency can be significantly reduced.

A novel presentation for channel selection problem in Brain-Computer Interfaces (BCI) is introduced here. Continuous presentation in a projected two-dimensional space of the Electroencephalograph (EEG) cap is proposed. A multi-objective particle swarm optimization method (D2MOPSO) is employed where particles move in the EEG cap space to locate the optimum set of solutions that minimize the number of selected channels and the classification error rate. This representation focuses on the local relationships among EEG channels as the physical location of the channels is explicitly represented in the search space avoiding picking up channels that are known to be uncorrelated with the mental task. In addition continuous presentation is a more natural way for problem solving in PSO framework. The method is validated on 10 subjects performing right-vs-left motor imagery BCI. The results are compared to these obtained using Sequential Floating Forward Search (SFFS) and shows significant enhancement in classification accuracy but most importantly in the distribution of the selected channels.

Bayesian Networks (BNs) are graphical probabilistic models that represent relationships that may exist between variables of a dataset. BN can be applied to data in a variety of different ways. Yet, using a BN requires knowing its structure. BN structure learning represents a challenge as the number of possible structures is very large. Search and score approaches have been used to address the problem. One of them, a Genetic Algorithm based on the K2 search (K2GA) has shown that BNs can be learned from many datasets. However, the computational cost which is involved is high while structures obtained from benchmark data often exhibit significant differences from known correct structures.

The Artificial Reaction Network (ARN) is a bio-inspired connectionist paradigm based on the emerging field of Cellular Intelligence. It has properties in common with both AI and Systems Biology techniques including Artificial Neural Networks, Petri Nets, and S-Systems. This paper discusses the temporal aspects of the ARN model using robotic gaits as an example and compares it with properties of Artificial Neural Networks. The comparison shows that the ARN based network has similar functionality.

The Artificial Reaction Network (ARN) is a bio-inspired connectionist paradigm based on the emerging field of Cellular Intelligence. It has properties in common with both AI and Systems Biology techniques including Artificial Neural Networks, Petri Nets, and S-Systems. In this paper, elements of temporal dynamics and pattern recognition are combined within a single ARN control system for a quadrupedal robot. The results show that the ARN has similar applicability to Artificial Neural Network models in robotic control tasks. In comparison to neural Central Pattern Generator models, the ARN can control gaits and offer reduced complexity. Furthermore, the results show that like spiky neural models, the ARN can combine pattern recognition and complex temporal control functionality in a single network.

Search and score techniques have been widely applied to the problem of learning Bayesian Networks (BNs) from data. Many implementations focus on finding an ordering of variables from which edges can be inferred. Although varying across data, most search spaces for such tasks exhibit many optima and plateaus. Such characteristics represent a trap for population-based algorithms as the diversity decreases and the search converges prematurely. In this paper, we study the impact of a distance mutation operator and propose a novel method using a population of agents that mutate their solutions according to their respective positions in the population. Experiments on a set of benchmark BNs confirm that diversity is maintained throughout the search. The proposed technique shows improvement on most of the datasets by obtaining BNs of similar of higher quality than those obtained by Genetic Algorithm methods.

In recent years, Markov Network EDAs have begun to find application to a range of important scientific and industrial problems. In this chapter we focus on several applications of Markov Network EDAs classified under the DEUM framework which estimates the overall distribution of fitness from a bitstring population. In Section 1 we briefly review the main features of the DEUM framework and highlight the principal features that havemotivated the selection of applications. Sections 2 - 5 describe four separate applications: chemotherapy optimisation; dynamic pricing; agricultural biocontrol; and case-based feature selection. In Section 6 we summarise the lessons learned from these applications. These include: comparisons with other techniques such as GA and Bayesian Network EDAs; trade-offs between modelling cost and reduction in search effort; and the use of MN models for surrogate evaluation.We also present guidelines for further applications and future research.

Fitness modelling is an area of research which has recently receivedmuch interest among the evolutionary computing community. Fitness models can improve the efficiency of optimisation through direct sampling to generate new solutions, guiding of traditional genetic operators or as surrogates for a noisy or long-running fitness functions. In this chapter we discuss the application of Markov networks to fitness modelling of black-box functions within evolutionary computation, accompanied by discussion on the relationship between Markov networks andWalsh analysis of fitness functions.We review alternative fitness modelling and approximation techniques and draw comparisons with the Markov network approach. We discuss the applicability of Markov networks as fitness surrogates which may be used for constructing guided operators or more general hybrid algorithms.We conclude with some observations and issues which arise from work conducted in this area so far.