Automated risk mitigation in business processes

This paper proposes a concrete approach for the automatic mitigation of risks that are detected during process enactment. Given a process model exposed to risks, e.g. a financial process exposed to the risk of approval fraud, we enact this process and as soon as the likelihood of the associated risk(s) is no longer tolerable, we generate a set of possible mitigation actions to reduce the risks' likelihood, ideally annulling the risks altogether. A mitigation action is a sequence of controlled changes applied to the running process instance, taking into account a snapshot of the process resources and data, and the current status of the system in which the process is executed. These actions are proposed as recommendations to help process administrators mitigate process-related risks as soon as they arise. The approach has been implemented in the YAWL environment and its performance evaluated. The results show that it is possible to mitigate process-related risks within a few minutes.

Cross-Lingual knowledge discovery : Chinese-to-English article linking in Wikipedia

In this paper we examine automated Chinese to English link discovery in Wikipedia and the effects of Chinese segmentation and Chinese to English translation on the hyperlink recommendation. Our experimental results show that the implemented link discovery framework can effectively recommend Chinese-to-English cross-lingual links. The techniques described here can assist bi-lingual users where a particular topic is not covered in Chinese, is not equally covered in both languages, or is biased in one language; as well as for language learning.

Walk@Work : an automated intervention to increase walking in university employees not achieving 10,000 daily steps

This study assessed the workday step counts of lower active (<10,000 daily steps) university employees using an automated, web-based walking intervention (Walk@Work). METHODS: Academic and administrative staff (n=390; 45.6±10.8years; BMI 27.2±5.5kg/m2; 290 women) at five campuses (Australia [x2], Canada, Northern Ireland and the United States), were given a pedometer, access to the website program (2010-11) and tasked with increasing workday walking by 1000 daily steps above baseline, every two weeks, over a six week period. Step count changes at four weeks post intervention were evaluated relative to campus and baseline walking. RESULTS: Across the sample, step counts significantly increased from baseline to post-intervention (1477 daily steps; p=0.001). Variations in increases were evident between campuses (largest difference of 870 daily steps; p=0.04) and for baseline activity status. Those least active at baseline (<5000 daily steps; n=125) increased step counts the most (1837 daily steps; p=0.001), whereas those most active (7500-9999 daily steps; n=79) increased the least (929 daily steps; p=0.001). CONCLUSIONS: Walk@Work increased workday walking by 25% in this sample overall. Increases occurred through an automated program, at campuses in different countries, and were most evident for those most in need of intervention.

Slice, mine and dice : complexity-aware automated discovery of business process models

Automated process discovery techniques aim at extracting models from information system logs in order to shed light into the business processes supported by these systems. Existing techniques in this space are effective when applied to relatively small or regular logs, but otherwise generate large and spaghetti-like models. In previous work, trace clustering has been applied in an attempt to reduce the size and complexity of automatically discovered process models. The idea is to split the log into clusters and to discover one model per cluster. The result is a collection of process models -- each one representing a variant of the business process -- as opposed to an all-encompassing model. Still, models produced in this way may exhibit unacceptably high complexity. In this setting, this paper presents a two-way divide-and-conquer process discovery technique, wherein the discovered process models are split on the one hand by variants and on the other hand hierarchically by means of subprocess extraction. The proposed technique allows users to set a desired bound for the complexity of the produced models. Experiments on real-life logs show that the technique produces collections of models that are up to 64% smaller than those extracted under the same complexity bounds by applying existing trace clustering techniques.

CADDS: an automated die design system for sheet-metal blanking

This article reports on the design and implementation of a Computer-Aided Die Design System (CADDS) for sheet-metal blanks. The system is designed by considering several factors, such as the complexity of blank geometry, reduction in scrap material, production requirements, availability of press equipment and standard parts, punch profile complexity, and tool elements manufacturing method. The interaction among these parameters and how they affect designers' decision patterns is described. The system is implemented by interfacing AutoCAD with the higher level languages FORTRAN 77 and AutoLISP. A database of standard die elements is created by parametric programming, which is an enhanced feature of AutoCAD. The greatest advantage achieved by the system is the rapid generation of the most efficient strip and die layouts, including information about the tool configuration.

Contemporaneous correlations within individuals in an automated web-based walking program (Walk@Work) over a six week period (three phases) of time

Introduction: Participants may respond to phases of a workplace walking program at different rates. This study evaluated the factors that contribute to the number of steps through phases of the program. The intervention was automated through a web-based program designed to increase workday walking. Methods: The study reviewed independent variable influences throughout phases I–III. A convenience sample of university workers (n=56; 43.6±1.7 years; BMI 27.44±.2.15 kg/m2; 48 female) were recruited at worksites in Australia. These workers were given a pedometer (Yamax SW 200) and access to the website program. For analyses, step counts entered by workers into the website were downloaded and mean workday steps were compared using a seemingly unrelated regression. This model was employed to capture the contemporaneous correlation within individuals in the study across observed time periods. Results: The model predicts that the 36 subjects with complete information took an average 7460 steps in the baseline two week period. After phase I, statistically significance increases in steps (from baseline) were explained by age, working status (full or part time), occupation (academic or professional), and self reported public transport (PT) use (marginally significant). Full time workers walked more than part time workers by about 440 steps, professionals walked about 300 steps more than academics, and PT users walked about 400 steps more than non-PT users. The ability to differentiate steps after two weeks among participants suggests a differential affect of the program after only two weeks. On average participants increased steps from week two to four by about 525 steps, but regular auto users had nearly 750 steps less than non-auto users at week four. The effect of age was diminished in the 4th week of observation and accounted for 34 steps per year of age. In phase III, discriminating between participants became more difficult, with only age effects differentiating their increase over baseline. The marginal effect of age by phase III compared to phase I, increased from 36 to 50, suggesting a 14 step per year increase from the 2nd to 6th week. Discussion: The findings suggest that participants responded to the program at different rates, with uniformity of effect achieved by the 6th week. Participants increased steps, however a tapering off occurred over time. Age played the most consistent role in predicting steps over the program. PT use was associated with increased step counts, while Auto use was associated with decreased step counts.

A multi-layered approach for site detection in UAS emergency landing scenarios using geometry-based image segmentation

This paper presents an alternative approach to image segmentation by using the spatial distribution of edge pixels as opposed to pixel intensities. The segmentation is achieved by a multi-layered approach and is intended to find suitable landing areas for an aircraft emergency landing. We combine standard techniques (edge detectors) with novel developed algorithms (line expansion and geometry test) to design an original segmentation algorithm. Our approach removes the dependency on environmental factors that traditionally influence lighting conditions, which in turn have negative impact on pixel-based segmentation techniques. We present test outcomes on realistic visual data collected from an aircraft, reporting on preliminary feedback about the performance of the detection. We demonstrate consistent performances over 97% detection rate.

GrabCutSFM : how 3D information improves unsupervised object segmentation

In this paper, we present an unsupervised graph cut based object segmentation method using 3D information provided by Structure from Motion (SFM), called Grab- CutSFM. Rather than focusing on the segmentation problem using a trained model or human intervention, our approach aims to achieve meaningful segmentation autonomously with direct application to vision based robotics. Generally, object (foreground) and background have certain discriminative geometric information in 3D space. By exploring the 3D information from multiple views, our proposed method can segment potential objects correctly and automatically compared to conventional unsupervised segmentation using only 2D visual cues. Experiments with real video data collected from indoor and outdoor environments verify the proposed approach.

Nonparametric semantic segmentation for 3D street scenes

In this paper we propose a method to generate a large scale and accurate dense 3D semantic map of street scenes. A dense 3D semantic model of the environment can significantly improve a number of robotic applications such as autonomous driving, navigation or localisation. Instead of using offline trained classifiers for semantic segmentation, our approach employs a data-driven, nonparametric method to parse scenes which easily scale to a large environment and generalise to different scenes. We use stereo image pairs collected from cameras mounted on a moving car to produce dense depth maps which are combined into a global 3D reconstruction using camera poses from stereo visual odometry. Simultaneously, 2D automatic semantic segmentation using a nonparametric scene parsing method is fused into the 3D model. Furthermore, the resultant 3D semantic model is improved with the consideration of moving objects in the scene. We demonstrate our method on the publicly available KITTI dataset and evaluate the performance against manually generated ground truth.

Automated marine mammal detection from aerial imagery

This paper presents two algorithms to automate the detection of marine species in aerial imagery. An algorithm from an initial pilot study is presented in which morphology operations and colour analysis formed the basis of its working principle. A second approach is presented in which saturation channel and histogram-based shape profiling were used. We report on performance for both algorithms using datasets collected from an unmanned aerial system at an altitude of 1000 ft. Early results have demonstrated recall values of 48.57% and 51.4%, and precision values of 4.01% and 4.97%.

Automated proofs of computational indistinguishability

We present a tool for automatic analysis of computational indistinguishability between two strings of information. This is designed as a generic tool for proving cryptographic security based on a formalism that provides computational soundness preservation. The tool has been implemented and tested successfully with several cryptographic schemes.

Automated aerial inspection guidance with improved turn planning

Aerial inspection of pipelines, powerlines, and other large linear infrastructure networks has emerged in a number of civilian remote sensing applications. Challenges relate to automating inspection flight for under-actuated aircraft with LiDAR/camera sensor constraints whilst subjected to wind disturbances. This paper presents new improved turn planning strategies with guidance suitable for automation of linear infrastructure inspection able to reduce inspection flight distance by including wind information. Simulation and experimental flight tests confirmed the flight distance saving, and the proposed guidance strategies exhibited good tracking performance in a range of wind conditions.

Validating a novel, low cost, automated malnutrition screening system as a predictor of nutritional risk in the Oncology Day Care Unit

Background Cancer-related malnutrition is associated with increased morbidity, poorer tolerance of treatment, decreased quality of life, increased hospital admissions, and increased health care costs (Isenring et al., 2013). This study’s aim was to determine whether a novel, automated screening system was a useful tool for nutrition screening when compared against a full nutrition assessment using the Patient-Generated Subjective Global Assessment (PG-SGA) tool. Methods A single site, observational, cross-sectional study was conducted in an outpatient oncology day care unit within a Queensland tertiary facility, with three hundred outpatients (51.7% male, mean age 58.6 ± 13.3 years). Eligibility criteria: ≥18 years, receiving anticancer treatment, able to provide written consent. Patients completed the Malnutrition Screening Tool (MST). Nutritional status was assessed using the PG-SGA. Data for the automated screening system was extracted from the pharmacy software program Charm. This included body mass index (BMI) and weight records dating back up to six months. Results The prevalence of malnutrition was 17%. Any weight loss over three to six weeks prior to the most recent weight record as identified by the automated screening system relative to malnutrition resulted in 56.52% sensitivity, 35.43% specificity, 13.68% positive predictive value, 81.82% negative predictive value. MST score 2 or greater was a stronger predictor of nutritional risk relative to PG-SGA classified malnutrition (70.59% sensitivity, 69.48% specificity, 32.14% positive predictive value, 92.02% negative predictive value). Conclusions Both the automated screening system and the MST fell short of the accepted professional standard for sensitivity (80%) or specificity (60%) when compared to the PG-SGA. However, although the MST remains a better predictor of malnutrition in this setting, uptake of this tool in the Oncology Day Care Unit remains challenging.

A data mining approach to improve the automated quality of data

This thesis describes the development of a robust and novel prototype to address the data quality problems that relate to the dimension of outlier data. It thoroughly investigates the associated problems with regards to detecting, assessing and determining the severity of the problem of outlier data; and proposes granule-mining based alternative techniques to significantly improve the effectiveness of mining and assessing outlier data.

Image segmentation and pose estimation of humans in video

This thesis introduces improved techniques towards automatically estimating the pose of humans from video. It examines a complete workflow to estimating pose, from the segmentation of the raw video stream to extract silhouettes, to using the silhouettes in order to determine the relative orientation of parts of the human body. The proposed segmentation algorithms have improved performance and reduced complexity, while the pose estimation shows superior accuracy during difficult cases of self occlusion.