Search

Containers such as Docker provide a lightweight virtualization technology. They have gained popularity in developing, deploying and managing applications in and across Cloud platforms. Container management and orchestration platforms such as Kubernetes run application containers in virtual clusters that abstract the overheads in managing the underlying infrastructures to simplify the deployment of container solutions. These platforms are well suited for modern web applications that can give rise to geographic fluctuations in use based on the location of users. Such fluctuations often require dynamic global deployment solutions. A key issue is to decide how to adapt the number and placement of clusters to maintain performance, whilst incurring minimum operating and adaptation costs. Manual decisions are naive and can give rise to: over-provisioning and hence cost issues; improper placement and performance issues, and/or unnecessary relocations resulting in adaptation issues. Elastic deployment solutions are essential to support automated and intelligent adaptation of container clusters in geographically distributed Clouds. In this article, we propose an approach that continuously makes elastic deployment plans aimed at optimizing cost and performance, even during adaptation processes, to meet service level objectives (SLOs) at lower costs. Meta-heuristics are used for cluster placement and adjustment. We conduct experiments on the Australia-wide National eResearch Collaboration Tools and Resources Research Cloud using Docker and Kubernetes. Results show that with only a 0.5 ms sacrifice in SLO for the 95th percentile of response times we are able to achieve up to 44.44% improvement (reduction) in cost compared to a naive over-provisioning deployment approach.

Animal pose estimation is important to understand animal behavior and injury. This is especially the case with horse lameness. In the past, studies have focused on combining trackable markers on animals with high-speed cameras to support gait detection. This is time consuming and often impractical. This paper explores the feasibility of using deep learning for markerless motion tracking. We trained a neural network to locate a horse's body parts, then analyzed whether the horse was lame based on their motion trajectory. Different animal pose estimation methods were studied using two state of the art convolutional network models. This paper presents the results of this work. A mobile application was also developed to test the practical application of marker-less recognition technology for detecting lameness in horses.

Object detection is a major task in computer vision. With the rapid development of machine learning in the past few decades, and more recently deep learning, it is now possible to utilise complex machine learning models to automatically detect and classify objects from potentially complex images. In this paper we consider machine (deep) learning networks suitable for detection and classification of (Australian) snakes and their deployment and performance in a mobile environment. We explore state of the art Convolutional Neural Networks (CNNs) and their use for transfer learning. We develop an iOS application supporting an offline (model-embedded on the device) approach and an online version where images are sent to a Cloud-based server for classification. We present the results and discuss the performance differences as well as the impact on the accuracy and time for classification for the two environments.

Deep learning for image recognition has received a lot of attention in recent years. In this paper we present a case study using two state-of-the-art deep learning libraries for image classification based on single phase (Single Shot Detection - SSD) and two-phase (Faster Region-based Convolutional Neural Network – Faster-RCNN) deep learning technologies. The case study is based on classification of lepidoptera: an order of species that includes butterflies and moths. We describe the data that was collected that underpinned this work. We also present the results and discuss the challenges with the work. Finally, we outline the implementation of a mobile application used as the client interface to the final solution.

Deepfakes allow for the automatic generation and creation of (fake) video content, e.g. through generative adversarial networks. Deepfake technology is a controversial technology with many wide reaching issues impacting society, e.g. election biasing. Much research has been devoted to developing detection methods to reduce the potential negative impact of deepfakes. Application of neural networks and deep learning is one approach. In this paper, we consider the deepfake detection technologies Xception and MobileNet as two approaches for classification tasks to automatically detect deepfake videos. We utilise training and evaluation datasets from FaceForensics++ comprising four datasets generated using four different and popular deepfake technologies. The results show high accuracy over all datasets with an accuracy varying between 91-98% depending on the deepfake technologies applied. We also developed a voting mechanism that can detect fake videos using the aggregation of all four methods instead of only one.

Machine learning has seen a significant surge and uptake across many diverse applications. The high flexibility, adaptability, and computing capabilities it provides extend traditional approaches used in multiple fields including network operation and management. Numerous surveys have explored machine learning algorithms in the context of networking, such as traffic engineering, performance optimization, and network security. Many machine learning approaches focus on clustering, classification, regression, and reinforcement learning. The innovation of this research, and the contribution of this paper lies in the detailed summary and comparison of learning-based congestion control approaches. Compared with traditional congestion control algorithms which are typically rule-based, capabilities to learn from historical experience are highly desirable. From the literature, it is observed that reinforcement learning is a crucial trend among learning-based congestion control algorithms. In this paper, we explore the performance of reinforcement learning-based congestion control algorithms and present current problems with reinforcement learning-based congestion control algorithms. Moreover, we outline challenges and trends related to learning-based congestion control algorithms.

Deep learning and neural networks have recently gained considerable attention and are now one of the most popular topics in modern computer science. One of the most promising applications of deep learning is in the field of computer vision and especially in the application of convolutional neural networks (CNNs) for object detection and classification of images. In this paper, we explore various CNN models to identify and classify common species of spiders found in Australia with specific focus on poisonous spiders. We compare the accuracy and performance of the deep learning models on a range of diverse spider species. We also develop an iOS application as the front-end user application.

Deep learning has gained much attention and been applied in many different fields. In this paper, we present a web application developed to identify and detect the number of distinct feral cats of Australia using deep learning algorithms targeted to data captured from a set of remote sensing cameras. Feral cat recognition is an especially challenging application of deep learning since the cats are often similar and, in some cases, differ only in very small patterns on their fur. Given the automated, sensor-based image capture from remote cameras, further challenges relate to the limited number of images available. To tackle this, we train four neural network models to distinguish 75 classes (i.e. distinct feral cats) using 30 to 80 images for each class. Based on a range of evaluation metrics, we select Mask R-CNN model with ImageNet pre-trained weights augmented with the ResNet-50 network as the basis for the web application. Using images of cats from cameras in five different forests around Victoria, we achieved an average accuracy of identification of individual (distinct) cats of 89.4% with a maximum accuracy of 96.3%. This work is used to support ecologists in the School of Biosciences at the University of Melbourne.

This paper describes the use of blockchain technology to ensure the integrity of data logs for a clinical registry, providing a technological means for a secure audit of that registry. The characteristics of a secure audit - tamper-resistance, verifiability, searchability and privacy - are described in their application to this registry, then an evaluation is performed detailing the use of blockchain to achieve these audit goals. The clinical registry tested - supporting ENSAT (the European Network for the Study of Adrenal Tumors) - is a production repository of clinical, phenotypic and genetic information about patients with adrenal cancer, a rare but often serious condition that affects approximately 1 in 600,000 of the world population. The registry is implemented using a standard n-tier web application, with a MySQL database back-end, and Java/JSP business logic and user interface. The information contributing to the full audit of data and usage of the registry, is captured in the application log-files, which are stored in two "mirrored" formats: ASCII text files compiled through the Java log4j project and in a MongoDB NoSQL database. Following a discussion of the relevant supporting features, the fully implemented solution - a private blockchain known as "ensatChain" - is evaluated for overall security, using the Microsoft "STRIDE" threat model.

An increasing amount of user location information is being generated due to the widespread use of social network applications and the ubiquitous adoption of mobile and wearable technologies. This data can be analysed to identify precise trajectories of individuals — where they went and when they were there. This is an obvious privacy issue, yet publication of real-time aggregate over such location streams can provide valuable resources for researchers and government agencies, e.g., in case of pandemics it would be very useful to identify who might have come into contact with an infected individual at a given time. Differential privacy techniques have become popular and widely adopted to address privacy concerns. However, there are three key issues that limit the application of existing differential privacy approaches to user trajectory data: (a) the heterogeneous nature of the trajectories, (b) uniform sliding window mechanisms do not meet individual privacy requirements and (c) limited privacy budgets and impact on data utility when applied to infinite data streams. To tackle these problems, this paper proposes a private real-time trajectory stream statistics publication mechanism utilizingdifferential privacy (DP-PSP). To relieve the heterogeneity issues, anchor point discovery (e.g., fixed locations like museums, parks, etc.) and road segmenting mechanisms are proposed. We provide an adaptive w-step sliding window approach that allows users to specify their own dynamic privacy budget distribution to optimize their own privacy budget. To preserve the data utility, we present multi-timestamp prediction models and private k-nearest neighbour selection and perturbation algorithms to reduce the amount of perturbation distortion induced through the differential privacy mechanism. Comprehensive experiments over real-life location-based social network user trajectories show that DP-PSP provides private data aggregate over infinite trajectory streams and boosts the utility and quality of the perturbed aggregation without compromising individual privacy.

A wide array of clouds have been built and adopted by business and research communities. Similarly, many research communities use workflow environments and tools such as Taverna and Galaxy to model the execution of tasks to support and expedite the reenactment of complex processes, and ultimately support the repeatability of science. As a result, a number of systems that integrate clouds with these workflow systems have emerged including CloudMap, CloudMan, Galaxy cloud, BioBlend, etc. Though these systems have proven to be successful in service and data integration, they have not dealt with the data security issue inherent in cloud-based systems with the outsourced models of infrastructure provisioning. For many domains, e.g., health, this poses serious challenges regarding the adoption of cloud infrastructures. Yet such domains also have much to gain from clouds especially given the explosion of genomics data and opportunities for personalized medicine in the big data era. This paper addresses this problem by presenting a trusted storage cloud for scientific workflows, called TruXy. The paper describes the TruXy architecture, the corresponding protocols and illustrates the adoption of TruXy to support collaborative bioinformatics research in endocrine genomics. A range of experiments have been performed to measure the performance of TruXy for processing of exome data sets on individuals with a rare genetic disorders: disorders of sex development (DSD). Our results show that the performance of TruXy is comparable to that of using a standalone workflow tool and that it can handle the big data security challenges required.

Road traffic congestion is an important issue in modern cities, however most existing traffic jam identification solutions are based on expensive facilities such as sensors or transport probe infrastructure with high deployment and management costs. As a result of the cost, such solutions are not ubiquitously deployed. The extensive use of smart mobile devices furnished with location-based competencies and the global popularity of microblogging applications like Twitter offers an opportunity to tackle these problems. Twitter users can serve as human traffic sensors providing real-time reflections of current traffic situations. However, these data can contain extensive personal privacy information that demands privacy preserving mechanisms for the user location and their current trajectory. Differential privacy can ensure that degrees of privacy in these trajectories can be preserved whilst allowing data analysis and mining of the Twitter content. This paper proposes an innovative private trajectories release model and associated algorithms with differential privacy guarantees that considers both data privacy and data utility. This includes development of a private reference system for calibrating separate (raw) users trajectories across obfuscated anchor points; construction of privacy supporting noise-enhanced prefix trees to release synthesis data privately, and comprehensive evaluation of both the accuracy and utility of the solutions in terms of a set of evaluation metrics based on real-life tweets-based user trajectories across the city of Melbourne.

Cloud computing allows to utilize servers in efficient and scalable ways through exploitation of virtualization technology. In the Infrastructure-as-a-Server (IaaS) Cloud model, many virtualized servers (instances) can be created on a single physical machine. There are many such Cloud providers that are now in widespread use offering such capabilities. However, Cloud computing has overheads and can constrain the scalability and flexibility, especially when diverse users with different needs wish to use the Cloud resources. To accommodate such communities, an alternative to Cloud computing and virtualization of whole servers that is gaining widespread adoption is micro-hosting services and container-based solutions. Container-based technologies such as Docker allow hosting of micro-services on Cloud infrastructures. These enable bundling of applications and data in a manner that allows their easy deployment and subsequent utilization. Docker is just one of the many such solutions that have been put forward. The purpose of this paper is to compare and contrast a range of existing container-based technologies for the Cloud and evaluate their pros and cons and overall performances. The OpenStack-based Australia-wide National eResearch Collaboration Tools and Resources (NeCTAR) Research Cloud (www.nectar.org.au) was used for this purpose. We describe the design of the experiments and benchmarks that were chosen and relate these to literature review findings.

Deep learning can bring significant improvements to a range of research areas and application domains. Computer vision is one of the key areas that can benefit from deep learning. In particular, object detection and classification can benefit from such approaches. In this work we focus on counting the number of images of individual fruit that can be used by fruit growers to estimate fruit crop yields. To achieve this, we apply two different state of the art convolutional neural networks (CNNs): Faster R-CNN and Single Shot Detection (SSD). CNNs depend on data for training and tuning of the models. In this paper we establish a dataset containing images for a range of fruit types. Using this data, we apply the models to identify the number of fruit in images and the challenges that are encountered. We present the experimental results of applying these approaches and illustrate their performance including the accuracy, time and loss when counting fruit on trees. We consider the future challenges in scaling this work to deal with more complex issues around fruit estimation at scale.

Deep learning is one of the most popular technologies and research areas in machine learning. Convolutional Neural Networks (CNNs) are a typical artificial neural network underpinning deep learning. They have been used in many fields including image recognition, natural language process, and through games such as AlphaGo. A CNN has many advantages such as efficient feature extraction, the simplicity of data format required and the small number of (hyper-)parameters that are required. This paper focuses on a particular application of deep learning: crowd counting. To address this, we apply a Multi-task, Cascaded Convolutional Neural Network (MTCNN). Compared to other models, this model has a good performance and requires a shorter inference time, with shallower network structure and smaller size. In order to demonstrate the value and feasibility of the technology and provide a friendly operating environment for users, the application was realised on both the iOS and Android platforms. A web platform was also developed to visualize the real-time data using a Firebase server.

Container-based virtualisation technologies are gaining more and more traction in recent years across Cloud platforms and this will likely continue in the coming years. As such, containers orchestration technologies are becoming indispensable. Kubernetes has become the de facto standard because of its robustness, maturity and rich features. To free users of the burden of having to configure and maintain complex Kubernetes infrastructures, but still make use of its functionalities, all major Cloud providers are now offering cloud-native managed Kubernetes alternatives. The goal of this paper is to investigate the performance of containers running in such hosted services. For this purpose, we conduct a series of experimental evaluations of containers to monitor the behaviour of system resources including CPU, memory, disk and network. A baseline consisting of a manually deployed Kubernetes cluster was built for comparison. In particular, we consider the Amazon Elastic Container Service for Kubernetes (EKS), Microsoft Azure Kubernetes Service (AKS) and Google Kubernetes Engine (GKE). The Australia-wide NeCTAR Research Cloud was used for the baseline.

Compared to the traditional approach of using virtual machines as the basis for the development and deployment of applications running in Cloud-based infrastructures, container technology provides developers with a higher degree of portability and availability, allowing developers to build and deploy their applications in a much more efficient and flexible manner. A number of tools have been proposed to orchestrate complex applications comprising multiple containers requiring continuous monitoring and management actions to meet application-oriented and non-functional requirements. Different container orchestration tools provide different features that incur different overheads. As such, it is not always easy for developers to choose the orchestration tool that will best suit their needs. In this paper we compare the benefits and overheads incurred by the most popular open source container orchestration tools currently available, namely: Kubernetes and Docker in Swarm mode. We undertake a number of benchmarking exercises from well-known benchmarking tools to evaluate the performance overheads of container orchestration tools and identify their pros and cons more generally. The results show that the overall performance of Kubernetes is slightly worse than that of Docker in Swarm mode. However, Docker in Swarm mode is not as flexible or powerful as Kubernetes in more complex situations.

With the prevalence of location-aware devices and applications, enormous volumes of human spatiotemporal trajectories are being produced. It is feasible to estimate the similarity between user movement patterns according to such trajectories, which can be regarded as a potential social tie between users. There are two key research challenges associated with social tie discovery from trajectories: (1) trajectories contain users' accurate locations and releasing such data for social tie discovery raises serious privacy concerns; (2) trajectories are archived as discrete approximations of actual movement patterns using different sampling strategies and rates which are intrinsically heterogeneous. To address these challenges, this paper proposes a Privacy-protected Social Tie Mining (P-STM) approach. It provides a new social tie discovery solution based on the similarity of calibrated trajectories incorporating three key components: (1) a location entropy-based indicative dense region (IDR) mining approach to handle the heterogeneity of trajectories under differential privacy; (2) a private model-based calibration system used to rewrite trajectories using a sanitized IDR set to improve the utility of sanitized trajectories for similarity evaluation; (3) a social tie mining approach to indicate potential social ties between individuals using the similarity trajectories, which aims at finding the acquaintances for users based on solely their local geographical activities. The proposed approach is evaluated using real-world trajectory datasets from location-based social networks.

This paper presents a novel application of machine learning through a mobile application that is used to address the requirements of hobby horticulturists through to the agricultural industry. Specifically, many large-scale farms such as fruit growers require information on the amount of chemicals, e.g. pesticides, to use on their crops. Hitherto, this is based on approximate estimates of the size of their fruit trees, where size here equates to the volume of their canopy and hence the amount of leaves that need to be sprayed. In this paper we present an approach to overcome such approximate measures through the development of a mobile phone-based application used to calculate the volume more accurately. For the hobby horticulturalist, the type of tree is also established using photographs of the tree and leaves.

Deep learning is one of the latest technologies in computer science. It allows using machines to solve problems in a manner similar to the human brain. Deep learning approaches have significantly improved the performance of visual recognition applications including image classification and image detection. In this paper we benchmark different deep learning models and present an Android application used to predict the location and breed of a given cat using a mobile phone camera. The average accuracy of the finalized model was 81.74%.

Argument component extraction is a challenging and complex high-level semantic extraction task. As such, it is both expensive to annotate (meaning training data is limited and low-resource by nature), and hard for current-generation deep learning methods to model. In this paper, we reevaluate the performance of state-of-the-art approaches in both single- and multi-task learning settings using combinations of character-level, GloVe, ELMo, and BERT encodings using standard BiLSTM-CRF encoders. We use evaluation metrics that are more consistent with evaluation practice in named entity recognition to understand how well current baselines address this challenge and compare their performance to lower-level semantic tasks such as CoNLL named entity recognition. We find that performance utilizing various pre-trained representations and training methodologies often leaves a lot to be desired as it currently stands, and suggest future pathways for improvement.

Big data technologies and a range of Government open data initiatives provide the basis for discovering new insights into cities; how they are planned, how they managed and the day-to-day challenges they face in health, transport and changing population profiles. The Australian Urban Research Infrastructure Network (AURIN --- www.aurin.org.au) project is one example of such a big data initiative that is currently running across Australia. AURIN provides a single gateway providing online (live) programmatic access to over 2000 data sets from over 70 major and typically definitive data-driven organizations across federal and State government, across industry and across academia. However whilst open (public) data is useful to bring data-driven intelligence to cities, more often than not, it is the data that is not-publicly accessible that is essential to understand city challenges and needs. Such sensitive (unit-level) data has unique requirements on access and usage to meet the privacy and confidentiality demands of the associated organizations. In this paper we highlight a novel geo-privacy supporting solution implemented as part of the AURIN project that provides seamless and secure access to individual (unit-level) data from the Department of Health in Victoria. We illustrate this solution across a range of typical city challenges in localized contexts around Melbourne. We show how unit level data can be combined with other data in a privacy-protecting manner. Unlike other secure data access and usage solutions that have been developed/deployed, the AURIN solution allows any researcher to access and use the data in a manner that meets all of the associated privacy and confidentiality concerns, without obliging them to obtain ethical approval or any other hurdles that are normally put in place on access to and use of sensitive data. This provides a paradigm shift in secure access to sensitive data with geospatial content.

Prediction of pollution is an increasingly important problem. It can impact individuals and their health, e.g. asthma patients can be greatly affected by air pollution. Traditional air pollution prediction methods have limitations. Machine learning provides one approach that can offer new opportunities for prediction of air pollution. There are however many different machine learning approaches and identifying the best one for the problem at hand is often challenging. In this paper air pollution data, specifically particulate matter of less than 2.5 micrometers (PM2.5) was collected from a variety of web-based resources and following, data cleansing analysed with different machine learning models including linear regression, Artificial Neural Networks and Long Short Term Memory recurrent neural networks. We consider the accuracy and the ability of these different models to predict unhealthy levels of pollution. The advantages and disadvantages of these models are also discussed.

The Australian Defence Forces (ADF) including the army, navy and air force often share resources. This is the case for helicopter training where the resources are often people, e.g. instructors. Understanding the current and future needs of the helicopter pilot training continuum is challenging. New students are continually entering the Defence services as trainees/cadets; passing or failing exams at specialist flying schools; progressing through or leaving the Defence services and potentially becoming trained pilots or instructors that can subsequently train future students. This continuum has an associated optimisation challenge. The ATHENA platform has been developed as a strategic discrete event simulation, optimisation and analysis system for manpower planning with specific focus on addressing the needs and demands of this helicopter pilot training continuum for the ADF. ATHENA has been developed as a simulation platform running on a Cloud infrastructure. This paper introduces ATHENA and describes the way in which the platform leverages container technologies to auto-scale across the Cloud with focus on the Kubernetes orchestration technology.

Deep learning provides the ability to train algorithms (models) that can tackle the problems of data classification and prediction based on deriving (learning) knowledge from raw data. Convolutional Neural Networks (CNNs) provides one commonly used approach for image classification and detection. In this work we describe a CNN-based method for detecting dogs in potentially complex images and subsequently consider the identification of the type/breed of dogs. The results achieve nearly 85% accuracy for breed classification for a set of 50 classes of dogs and 64% accuracy for 120 other less common dog types. An iOS application and associated big data processing infrastructure utilizing a variety of GPUs was used to support the image classification algorithms.

Due to the prevalence of location-based services, it has now become possible to infer social connections between people by observing their spatial behaviors over time. Such spatial behaviors, if shared, can be utilized to identify and recommend friends for web-based social service users. However, such approaches cannot be implemented without solving two key challenges: (a) guaranteeing an individual's privacy in the shared spatiotemporal data, and (b) addressing the inherent sparseness of shared spatiotemporal data. In this paper, we propose a Privacy-Aware Identification and Recommendation of Spatio-Friends (PAIRS) approach, that can infer and recommend potential social connections by analyzing spatiotemporal information of social media users using robust privacy guarantee mechanisms. To achieve this, PAIRS constructs co-occurrence profiles using a cluster-based anchor representation to alleviate the sparseness of shared spatiotemporal information. It utilizes the diversity, time and weighted frequency-based inference to efficiently infer the strength of potential social connections from co-occurrence profile by reducing the negative impact of coincidences and thereby enhances accuracy. To tackle the privacy concerns, PAIRS sanitizes the cluster-based anchors, the location entropy values as well as the co-occurrence profile under differential privacy, including optimization mechanisms to handle trade-offs in utility and privacy. Extensive experiments are conducted with real-world datasets including both individuals' spatiotemporal data and their actual social connections. We confirm that our approach can achieve two often contradictory goals: a provable robust privacy protection for sharing data and an efficient social strength inference and spatio-friend identification mechanism. Specifically, PAIRS remains approximately 70% accuracy (precision) and 80% efficiency (recommendation potential) after perturbation.

Real-time data processing is often a necessity as it can provide insights that have less value if discovered off-line or after the fact. However, large-scale stream processing systems are non-trivial to build and deploy. While there are many frameworks that allow users to create large-scale distributed systems, there remains many challenges in understanding the performance, cost of deployment and considerations and impact of potential (partial) outages on real-time systems performance. Our work considers the performance of Cloud-based stream processing systems in terms of back-pressure and expected utilization. The performance of an exemplar stream application is explored using different Cloud-based virtual machine resources and where the scale of deployment and cost benefits are taken into consideration in relation to the overall performance. To achieve this, we develop an algorithm based on queueing theory to predict the throughput and latency of stream data processing while supporting system stability. Our methodology for making fundamental measurements is applicable to mainstream stream processing frameworks such as Apache Storm and Heron. The method is especially suitable for large-scale distributed stream processing where jobs can run for extended time periods. We benchmark the performance of the system on the national research cloud of Australia (Nectar), and present a performance analysis based on estimating the overall effective utilization.

Clustering algorithms are essential for many big data applications involving point-based data, e.g. user generated social media data from platforms such as Twitter. One of the most common approaches for clustering is DBSCAN. However, DBSCAN has numerous limitations. The algorithm itself is based on traversing the whole dataset and identifying the neighbours around each point. This approach is not suitable when data is created and streamed in real-time however. Instead a more dynamic approach is required. This paper presents a new approach, RT-DBSCAN, that supports real-time clustering of data based on continuous cluster checkpointing. This approach overcomes many of the issues of existing clustering algorithms such as DBSCAN. The platform is realised using Apache Spark running over large-scale Cloud resources and container based technologies to support scaling. We benchmark the work using streamed social media content (Twitter) and show the advantages in performance and flexibility of RT-DBSCAN over other clustering approaches.

This paper describes a warrant classification system for SemEval 2018 Task 12, that attempts to learn semantic representations of reasons, claims and warrants. The system consists of 3 stacked LSTMs: one for the reason, one for the claim, and one shared Siamese Network for the 2 candidate warrants. Our main contribution is to force the embeddings into a shared feature space using vector operations, semantic similarity classification, Siamese networks, and multi-task learning. In doing so, we learn a form of generative implication, in encoding implication interrelationships between reasons, claims, and the associated correct and incorrect warrants. We augment the limited data in the task further by utilizing WordNet synonym “fuzzing”. When applied to SemEval 2018 Task 12, our system performs well on the development data, and officially ranked 8th among 21 teams.

Semantic annotation framework that allows enriching locations or trajectories with semantic abstractions of the raw spatiotemporal data benefits understanding the semantic behavior of moving objects. Existing semantic annotation approaches mainly analyze specific parts of a trajectory, e.g. stops, in association with data from 3rd party geographic sources, e.g. (POI) points-of-interest, road networks. However, these semantic resources are static thus miss important dynamic event information. Recent location-based social networking provides a new dynamic and prevalent source of human activity data that can be a potential semantic resource for annotation. However, using the large-scale spatiotemporal data from online social media gives rise to privacy concerns. This paper thus presents a privacy-preserving semantic annotation framework P-SAFE that (i) identifies dynamic region of interest (DRI) from large-scale data provided by location based social networks whilst labelling of DRI into appropriate categories derived from spatial and temporal features of geotags, (ii) aligns trajectories to a set of DRI and enriches trajectories with semantics annotation derived from aligned DRI via THMM model, and (iii) embeds robust privacy-preserving mechanisms under differential privacy in each stage that accesses to raw data. P-SAFE approach tackles the privacy and utility trade-offs for meaningful geographic regions identification and labeling as well as trajectory semantic annotation under differential privacy whilst combining them into a single task. We demonstrate the effectiveness of P-SAFE approach on a dataset of large-scale geotagged tweets and a benchmark trajectory dataset for DRI construction and trajectory semantic annotation evaluation. The experimental results illustrate that P-SAFE not only provides robust privacy guarantees but remains approximate 45–56% accuracy for meaningful geographic regions labelling and 62–76% accuracy for trajectory semantic annotation.

Niemann-Pick Types A and B (NPA/B) diseases are autosomal recessive lysosomal storage disorders caused by the deficient activity of acid sphingomyelinase (ASM) because of the mutations in the SMPD1 gene. Here, we provide a comprehensive updated review of already reported and newly identified SMPD1 variants. Among them, 185 have been found in NPA/B patients. Disease-causing variants are equally distributed along the SMPD1 gene; most of them are missense (65.4%) or frameshift (19%) mutations. The most frequently reported mutation worldwide is the p.R610del, clearly associated with an attenuated NP disease type B phenotype. The available information about the impact of 52 SMPD1 variants on ASM mRNA and/or enzymatic activity has been collected and whenever possible, phenotype/genotype correlations were established. In addition, we created a locus-specific database easily accessible at http://www.inpdr.org/genes that catalogs the 417 SMPD1 variants reported to date and provides data on their in silico predicted effects on ASM protein function or mRNA splicing. The information reviewed in this article, providing new insights into the genotype/phenotype correlation, is extremely valuable to facilitate diagnosis and genetic counseling of families affected by NPA/B.

With the explosive growth of location-aware devices and global adoption of social network applications, enormous volumes of spatiotemporal data are being produced. These can be perceived as gazetteers that record frequently visited locations, e.g. shopping malls and museums, and potentially more sensitive locations, e.g. an individual's home/work locations. Density-based clustering approaches are generally used for gazetteer discovery. However, existing clustering solutions are inefficient for big data scenarios and often disregard mobility features derived from trajectories data. Further, automated gazetteer discovery applications may cause privacy concerns. In this paper, we propose a sensitive gazetteer automated discovery approach based on Ω-cluster with robust privacy controls. The approach identifies sensitive gazetteers from massive trajectory data, with location entropy-based filtering used to reduce the number of uninteresting clusters whilst considering mobility features of trajectories. A parallelized solution is implemented to scale across the cloud using memory-oriented data processing solutions based upon Apache Spark. We embed this algorithm in a privacy-preserving mechanism and subsequently release sanitized gazetteers. Through extensive experiments using synthetic and real trajectory datasets from the location based social network (Twitter), we demonstrate the effectiveness and efficiency of our approach.

It is a simple task for humans to visually identify objects. However, computer-based image recognition remains challenging. In this paper we describe an approach for image recognition with specific focus on automated recognition of plants and flowers. The approach taken utilizes deep learning capabilities and unlike other approaches that focus on static images for feature classification, we utilize video data that compensates for the information that would otherwise be lost when comparing a static image with many others images of plants and flowers. We describe the steps taken in data collection, data cleaning and data purification, and the deep learning algorithms that were subsequently applied. We describe the mobile (iOS) application that was designed and finally we present the overall results that show that in the work undertaken thus far, the approach is able to identify 122/125 plants and 47/50 genera selected with degrees of confidence up to 95%. We also describe the performance speed up through the use of Cloud-based resources.

Background Computational bioinformatics workflows are extensively used to analyse genomics data, with different approaches available to support implementation and execution of these workflows. Reproducibility is one of the core principles for any scientific workflow and remains a challenge, which is not fully addressed. This is due to incomplete understanding of reproducibility requirements and assumptions of workflow definition approaches. Provenance information should be tracked and used to capture all these requirements supporting reusability of existing workflows. Results We have implemented a complex but widely deployed bioinformatics workflow using three representative approaches to workflow definition and execution. Through implementation, we identified assumptions implicit in these approaches that ultimately produce insufficient documentation of workflow requirements resulting in failed execution of the workflow. This study proposes a set of recommendations that aims to mitigate these assumptions and guides the scientific community to accomplish reproducible science, hence addressing reproducibility crisis. Conclusions Reproducing, adapting or even repeating a bioinformatics workflow in any environment requires substantial technical knowledge of the workflow execution environment, resolving analysis assumptions and rigorous compliance with reproducibility requirements. Towards these goals, we propose conclusive recommendations that along with an explicit declaration of workflow specification would result in enhanced reproducibility of computational genomic analyses. Electronic supplementary material The online version of this article (doi:10.1186/s12859-017-1747-0) contains supplementary material, which is available to authorized users.

Software Defined Networking (SDN) enables network innovation and brings flexibility by separation of the control and data planes and logically centralized control. However, this network paradigm complicates flow rule management. Current approaches generally install rules reactively after table misses or pre-installs them by flow prediction. Such approaches consume nontrivial network resources during interactions between the controller and switches (especially for maintaining consistency). In this paper, we explore an intelligent rule management scheme (IRMS), which extends the one-big-switch model and employs a hybrid rule management approach. To achieve this, we first transform all rules into path-based and node-based rules. Path-based rules are pre-installed whilst the paths for flows are selected at the edge switches of the network. To maintain consistency of forwarding paths, we update path-based rules as a whole and employ a lazy update policy. Node-based rules are optimally partitioned into disjoint chunks by an intelligent partition algorithm and organized hierarchically in the flow table. In this way, we significantly reduce the interaction cost between the control and data planes. This scheme enforces an efficient sliding window policy to enhance the hit rate for the installed chunks. We evaluate our scheme by comprehensive experiments. The results show that IRMS reduces the total flow entries by more than 71% on average and the update time by over 56%. IRMS also reduces the flow setup requests by more than one order of magnitude.

Estimation of crowd sizes or the occupancy of buildings and skyscrapers can often be essential. However, traditional ways of estimation through manual counting, image processing or in the case of skyscrapers, through total water usage are awkward, inefficient and often inaccurate. Social media has developed rapidly in the last decade. In this work, we provide novel solutions to estimate the population of suburbs and skyscrapers—so-called micro-populations, through the use of social media. We develop a big data solution leveraging large-scale harvesting and analysis of Twitter data. By harvesting real-time tweets and clustering tweets within suburbs and skyscrapers, we show how micro-populations can be calculated. To validate this, we construct linear and spatial models for the suburbs in four cities of Australia using census data and geospatial data models (shapefiles). Our prediction of micro-population shows that Twitter can indeed be used for population prediction with a high degree of accuracy.

Nowadays, more and more data are being generated and collected. This is especially so in our cities. Big data as a popular research topic has brought about a range of new challenges that must to be tackled to support both commercial and research demands. The transport area is one example that has much to benefit from big data capabilities in allowing to process voluminous amounts of data that is created in real time and in vast quantities. Tackling these big data issues requires capabilities not typically found in common Cloud platforms. The high speed of production of data (velocity) of high volume data demands a distributed file system for storing and duplicating data; high performance computing engines capable of processing such large quantities of data in real time; a reliable database system able to optimize the indexing and querying of the data, and essentially a flexible implementation for scaling the usage of computation resources based on the bursty nature of data production. In this paper, we present and benchmark SMASH, a generic and highly scalable Cloud-based solution. We focus here on the implementation and especially on the utilization of the SMASH software stack and how it can be used to process large scale traffic data - although we note that the solution can be applied to other high velocity big data processing areas.

Clustering techniques can group genes based on similarity in biological functions. However, the drawback of using clustering techniques is the inability to identify an optimal number of potential clusters beforehand. Several existing optimization techniques can address the issue. Besides, clustering validation can predict the possible number of potential clusters and hence increase the chances of identifying biologically informative genes. This paper reviews and provides examples of existing methods for clustering genes, optimization of the objective function, and clustering validation. Clustering techniques can be categorized into partitioning, hierarchical, grid-based, and density-based techniques. We also highlight the advantages and the disadvantages of each category. To optimize the objective function, here we introduce the swarm intelligence technique and compare the performances of other methods. Moreover, we discuss the differences of measurements between internal and external criteria to validate a cluster quality. We also investigate the performance of several clustering techniques by applying them on a leukemia dataset. The results show that grid-based clustering techniques provide better classification accuracy; however, partitioning clustering techniques are superior in identifying prognostic markers of leukemia. Therefore, this review suggests combining clustering techniques such as CLIQUE and k-means to yield high-quality gene clusters.

The value that can be extracted from big data greatly motivates organizations to explore data analytics technologies for better decision making and problem solving in a wide range of application domains. Cloud computing greatly eases and benefits big data analytics by offering on-demand and scalable computing infrastructures, platforms, and applications as services. Big data Analytics-as-a-Service (AaaS) platforms aim to deliver data analytics as consumable services in cloud computing environments in a pay as you go model with Service Level Agreement (SLA) guarantees. Resource scheduling for AaaS platforms is significant as big data analytics requires large-scale computing, which can consume huge amounts of resources and incur high resource costs. Our research focuses on proposing automatic and scalable resource scheduling algorithms to maximize the profits for AaaS platforms while delivering AaaS services to users with SLA guarantees on budgets and deadlines to allow timely responses with controllable costs. In this paper, we model and formulate the profit optimization resource scheduling problem and propose an optimization scheduling algorithm that maximizes profits for AaaS platforms and guarantees SLAs for query requests. Experimental evaluations show that the profit optimization scheduling algorithm performs significantly better in cost saving and profit enhancement compared to the state-of-the-art scheduling algorithms.

Unprecedented volumes of location-based information have been produced as a result of the widespread adoption of social network applications and GPS-enabled devices and sensors. Publication of such location data can provide valuable resources for researchers and government agencies in applications ranging from near real-time population-wide health monitoring to planning for future cities. However, such data hold personally identifying information, which gives rise to many privacy issues. There is thus a pressing need for ways to restrict this inherently identifying location-related information, however ideally we would like to preserve the utility of the data. Importantly, any such solution has to be scalable to large population-wide data scenarios. To tackle this, we introduce a novel differentially private hierarchical location sanitization (DPHLS) approach based on the concept “(α, r)-dataset” implemented through a Variable Order Mobility Markov Model (VO3M). We show how this system allows individual locations in personal trajectories to be protected using selection and frequency perturbation mechanisms using the “(α, r)-dataset”, leveraging past (published) location histories to obfuscate the user location in a flexible and controllable manner. The effectiveness and efficiency of the proposed solution is evaluated through the big data experiments that have been carried out using an OpenStack-based Cloud and Apache Sparkbased platform utilising large-scale social media trajectories. The experimental results suggest that the privacy publication algorithm can successfully scale to big data scenarios whilst retaining the utility of the datasets (trajectories) and preserving individual user privacy.

With the global adoption of smart mobile devices equipped with localization capabilities and broad popularity of microblogging facilities like Twitter, the need for personal privacy has never been greater. This is especially so with computational and data processing infrastructures such as clouds that support big data analysis. Differential privacy of geospatially tagged data such as tweets can potentially ensure that degrees of location privacy can be preserved while allowing the information (tweet contents) to be used for research and analysis, e.g., sentiment analysis. In this paper, we evaluate differential location pattern-mining approaches considering both privacy and precision of geo-located tweets clustered according to Geo-Locations of Interest (GLI). We consider both the privacy protection strength and the accuracy of results, measuring the Euclidean distance between centroids of real GLIs and obfuscated ones, i.e., those incorporating privacy-preserving noise. We record the performance and sensitivity of the approach. We show how privacy and location precision are trade-offs, i.e., the higher the degree of privacy protection, the fewer the GLIs will be identified. We also quantify these trade-offs and their associated sensitivity levels. We illustrate the work through a big data case study on use of Twitter data for traffic-related data protection.

Protecting (authorizing) access to individual web services has been explored in many research efforts. The focus of such research is to ensure that authorized users with appropriate credentials are able to access resources under controlled and authorized security. However, integrating and/or composing such services, e.g. through workflow environments in collaborative environments, remains an open challenge. A key issue is to ensure the security requirements of each individual service and hence their service provider, whilst still allowing them to be used as part of a workflow execution – potentially by users with diverse security credentials. The aim of this research is to provide a security-oriented workflow framework that supports such secure workflow enactment. The framework supports a variety of security evaluation mechanisms using diverse security credential aggregation mechanisms leveraging centralized and decentralized security with push/pull of security information as required. We describe the core components of this framework and how they can be used to support security-oriented workflow enactment.

Decentralized enactment of workflows is generally advocated for data intensive scientific applications. This approach offers a number of advantages including avoiding a single (centralized) point of failure, and associated (centralized) performance bottlenecks. However, such services are often assumed to be openly available with little or no security. This is increasingly not the case. Instead organizations and the services they offer require autonomous ways of defining and enforcing their own access control policies supporting fine-grained authorization. Tackling such scenarios raises many issues regarding workflow definition, enactment and potential re-enactment. This research explores the issues related to security-oriented decentralized workflow definition and enactment. The solution is respectful of the need for autonomy of service providers and allows each participating service to define their own access control policies. We also show how different security delivery models can be supported leveraging the pulling and pushing of security credentials.

New technology supporting large-scale (big) data collection and associated data analytics is challenging personal privacy in unprecedented and unforeseen ways. Big Data is typically high volume, high velocity, heterogeneous, and distributed with varying degrees of veracity. Big Data can be created and collected by individuals, organizations, or external agencies, often with the aim of applying data analytics to improve services, products, or decision-making functions that can potentially add competitive advantages. The tools and infrastructure for supporting the capture and analysis of Big Data are widespread, for example, MapReduce and ElasticSearch are algorithms now taught at the undergraduate level that allow processing of extensive and heterogeneous data. Data analytics is increasingly being used to derive meaningful patterns from such data. However such approaches have a potential downside with regard to privacy. Social media is one example of Big Data that raises many challenges in terms of an individual’s privacy. Many consider that the privacy issues for data that is randomly posted by individuals in tweets, blogs, and through images onto the Internet are less important; however, many social media data sets contain extensive data that can reveal information that users may be unaware of, including their geospatial location. In this chapter we explore these issues, focusing especially on the challenges of privacy, with a specific focus on location privacy. We outline how Cloud infrastructures and related technologies can be used to explore and highlight the issues that can arise with the privacy of locations and the consequences of this on an individual’s privacy.

Many biomedical research collaborations are focused on the establishment of web-based databases that capture phenotypic and in some cases genotypic information, targeted to specific diseases – so called disease registries. Such resources are often used for clinical matchmaking and allow information on patients and patient disorders to be shared by clinicians with wider biomedical research communities outside of a given hospital setting, and potentially with patients and/or patient advisory groups. Whilst addressing aspects of clinical collaborations through making targeted biomedical data accessible - such registries are really only a starting point for what can be achieved to support research collaborations. In particular, registries should ideally be augmented with a portfolio of additional service offerings that facilitate secure research collaborations: bio-banking and bio-sample data tracking capabilities; support for feasibility analysis on clinical trials and studies; offer seamless data transfer to/from clinical trials; provide search and analytical capabilities in a user-driven research environment. Such a feature rich, Internet-based virtual research environment (VRE) has been established as part of the European Union funded ENS@T-CANCER (http://www.ensat-canser.eu) project that has a particular focus on supporting research into four primary types of adrenal tumors. This paper provides an overview of the ENS@T-CANCER VRE, outlining its core capabilities and how it has galvanized previously and largely fragmented, country-specific database and registry efforts. To illustrate the utility of the ENS@T-CANCER VRE for clinical trials we focus on the full 4-phase, international, multi-centre Prospective Monoamine Tumor (PMT) clinical trial and its utilisation of capabilities provided by the VRE.

Scientific experiments often involve use of shared resources across organization boundaries in distributed collaborative environments. They are more often enabled through web services. A plethora of research is undertaken to protect individual web services. They include centralized security models wherein the main focus is on centralised Virtual Organization (VO) specific attribute authorities, e.g. VOMS, which can be used by collaborative service providers to make authorisation decisions. And a decentralized security model wherein each service provider themselves are responsible for the assignment of roles/privileges to the different members of collaborative environments. Workflows themselves can be orchestrated in centralized or decentralized orchestration models. In this research work we have identified a number of architectural design patterns for security-enabled workflows executions. These patterns are based on the different workflows execution and security models. The key issues in such patterns as well as a rationale of choice are provided. An overview of a security-oriented workflow framework is provided that can tackle some of the issues identified in these patterns.

Grid computing facilitates resource sharing typically to support distributed virtual organizations (VO). The multi-institutional nature of a grid environment introduces challenging security issues, especially with regard to authentication and authorization. This article presents a state-of-the-art review of major grid authentication and authorization technologies. In particular we focus upon the Internet2 Shibboleth technologies and their use to support federated authentication and authorization to support interinstitutional sharing of remote grid resources that are subject to access control. We outline the architecture, features, advantages, limitations, projects, and applications of Shibboleth in a grid environment. The evidence suggests that Shibboleth meets many of the demands of the research community in accessing and using grid resources.

The UK Department of Trade and Industry (DTI) funded BRIDGES project (Biomedical Research Informatics Delivered by Grid Enabled Services) has developed a Grid infrastructure to support cardiovascular research. This includes the provision of a compute Grid and a data Grid infrastructure with security at its heart. In this paper we focus on the BRIDGES data Grid. A primary aim of the BRIDGES data Grid is to help control the complexity in access to and integration of a myriad of genomic data sets through simple Grid-based tools. We outline these tools, how they are delivered to the end user scientists. We also describe how these tools are to be extended in the BBSRC funded Grid Enabled Microarray Expression Profile Search (GEMEPS) to support a richer vocabulary of search capabilities to support mining of microarray data sets. As with BRIDGES, fine grain Grid security underpins GEMEPS.

In recent times, big data has become a popular research topic and brought about a range of new challenges that must be tackled to support many commercial and research demands. The transport arena is one example that has much to benefit from big data capabilities in allowing to process voluminous amounts of data that is created in real time and in vast quantities. Tackling these big data issues requires capabilities not typically found in common Cloud platforms. This includes a distributed file system for capturing and storing data, a high performance computing engine able to process such large quantities of data, a reliable database system able to optimize the indexing and querying of the data, and geospatial capabilities to visualize the resultant analyzed data. In this paper we present SMASH, a generic and highly scalable Cloud-based architecture and its implementation that meets these many demands. We focus here specifically on the utilization of the SMASH software stack to process large scale traffic data for Adelaide and Victoria although we note that the solution can be applied to other big data processing areas. We provide performance results on SMASH and compare it with other big data solutions that have been developed.

Twitter, the popular microblogging platform, has more than five hundred million registered users -- so called Tweeters. These Tweeters generate a large amount of information every day. This information, whilst noisy and diverse is so voluminous that it can provide a key source of information that can support a range of research activities. Transport and traffic accidents are the examples chosen here. At the same time, Melbourne has been named the world's most liveable city, but this city continues to face challenges caused by the continuous population growth and the issues facing the transport systems caused by increased traffic volumes. This has implications for road safety. The Victorian government has implemented a transportation accident black spot programme however this information is largely historic in nature and does not capture real time traffic accidents. We have designed a Cloud-based software system that correlates historic accident black spot information with Twitter and uses this to benchmark the accuracy of Twitter for identifying and verifying the location of traffic accident black spots. The ultimate goal is to use Twitter data to identify accidents in real-time with a measurable degree of confidence.