Search

The application of big data, radiomics, machine learning, and artificial intelligence (AI) algorithms in radiology requires access to large data sets containing personal health information. Because machine learning projects often require collaboration between different sites or data transfer to a third party, precautions are required to safeguard patient privacy. Safety measures are required to prevent inadvertent access to and transfer of identifiable information. The Canadian Association of Radiologists (CAR) is the national voice of radiology committed to promoting the highest standards in patient-centered imaging, lifelong learning, and research. The CAR has created an AI Ethical and Legal standing committee with the mandate to guide the medical imaging community in terms of best practices in data management, access to health care data, de-identification, and accountability practices. Part 2 of this article will inform CAR members on the practical aspects of medical imaging de-identification, strengths and limitations of de-identification approaches, list of de-identification software and tools available, and perspectives on future directions. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) software that analyzes medical images is becoming increasingly prevalent. Unlike earlier generations of AI software, which relied on expert knowledge to identify imaging features, machine learning approaches automatically learn to recognize these features. However, the promise of accurate personalized medicine can only be fulfilled with access to large quantities of medical data from patients. This data could be used for purposes such as predicting disease, diagnosis, treatment optimization, and prognostication. Radiology is positioned to lead development and implementation of AI algorithms and to manage the associated ethical and legal challenges. This white paper from the Canadian Association of Radiologists provides a framework for study of the legal and ethical issues related to AI in medical imaging, related to patient data (privacy, confidentiality, ownership, and sharing); algorithms (levels of autonomy, liability, and jurisprudence); practice (best practices and current legal framework); and finally, opportunities in AI from the perspective of a universal health care system. [ABSTRACT FROM AUTHOR]

Since individual neural networks have limited deep expressiveness and effectiveness, many learning frameworks face difficulties in the availability and balance of sample selection. As a result, in change detection, it is difficult to upgrade the hit rate of a high-performance model on both positive and negative pixels. Therefore, supposing that the sacrificed components coincide perfectly with the important evaluation objectives, such as positives, it would lose more than gain. To address this issue, in this paper, we propose a multi-visual collaborative deep network (MV-CDN) served by three collaborative network members that consists of three subdivision approaches, the CDN with one collaborator (CDN-C), CDN with two collaborators (CDN-2C), and CDN with three collaborators (CDN-3C). The purpose of the collaborator is to re-evaluate the feature elements in the network transmission, and thus to translate the group-thinking into a more robust field of vision. We use three sets of public double-temporal hyperspectral images taken by the AVIRIS and HYPERION sensors to show the feasibility of the proposed schema. The comparison results have confirmed that our proposed schema outperforms the existing state-of-the-art algorithms on the three tested datasets, which demonstrates the broad adaptability and progressiveness of the proposal. [ABSTRACT FROM AUTHOR]

Background: Blended online courses, which combine synchronous and asynchronous online activities, have expanded rapidly in higher education. How to enhance student engagement in such courses is unclear, although it is recognized that student engagement is malleable through instructional strategies. Objectives: Given the above, this study aims to examine the influence of categories of strategies on student engagement in blended online courses. Methods: A conceptual framework of instructional strategies indicated as fostering student engagement in the relevant literature was first presented, divided in eight categories (structure, pace, relevance, active, choice, relationships, explanations, guide). Then a research framework linking the categories of strategies to student engagement dimensions (emotional‐cognitive, social, behavioral) was built and tested in blended online courses. Data collected in various disciplines and university levels at four universities (n = 482) were examined using partial least squares structural equation modeling. Results and Conclusions: The structural model examination confirmed the combined effects of categories of instructional strategies on student engagement in such courses in all disciplines. Particularly, this study revealed that 1) establishing trusting relationships, 2) demonstrating the relevance of activities, content, and resources, and 3) maintaining a sustained course pace significantly impacted student engagement in blended online courses in all disciplines. Takeaways: This study draws upon the blended learning literature to bring together key instructional strategies that foster student engagement while highlighting empirical quantitative evidence of their effects on student engagement in blended online courses. Detailed measures of categories of instructional strategies and student engagement dimensions also provide reliable instruments for future research. Lay Description: What is currently known about the subject matter?: Student engagement has important repercussions on perseverance, in‐depth learning, satisfaction, and academic success.Blended online courses merge the benefits of synchronous interactions and online flexibility in terms of time, place, or even study pace.Student engagement is influenced by contextual variations like course modalities and instructional strategies.The literature lacks a quantitative large‐scale portrait of how instructional strategies influence student engagement in blended online courses in higher education. What does this paper add to the existing body of research?: Partial least squares structural equation modelling was used to examine the effects of eight categories of instructional strategies on emotional‐cognitive, social, and behavioural dimensions of student engagement.Emotional‐cognitive engagement is largely affected by instructional strategies, while social and behavioural engagement are moderately affected by these.Establishing trusting relationships; demonstrating the relevance of activities, content, and resources; and maintaining a sustained course pace have the most significant effects on student engagement in blended online courses. Implications of study findings for practitioners: Teachers should explicitly support students in the process of developing interpersonal relationships in blended online courses to foster student social engagement, that is, interactions and sense of connectedness between them.Ensuring a sustained pace, dynamism, and a diversity of teaching and learning activities, both synchronously and asynchronously, promotes student engagement in blended online courses.Highlighting or demonstrating the relevance of activities and contents fosters emotional‐cognitive and behavioural student engagement. [ABSTRACT FROM AUTHOR]

This paper investigates how to best compare algorithms for predicting chronic homelessness for the purpose of identifying good candidates for housing programs. Predictive methods can rapidly refer potentially chronic shelter users to housing but also sometimes incorrectly identify individuals who will not become chronic (false positives). We use shelter access histories to demonstrate that these false positives are often still good candidates for housing. Using this approach, we compare a simple threshold method for predicting chronic homelessness to the more complex logistic regression and neural network algorithms. While traditional binary classification performance metrics show that the machine learning algorithms perform better than the threshold technique, an examination of the shelter access histories of the cohorts identified by the three algorithms show that they select groups with very similar characteristics. This has important implications for resource constrained not-for-profit organizations since the threshold technique can be implemented using much simpler information technology infrastructure than the machine learning algorithms. [ABSTRACT FROM AUTHOR]

The Connected Vertex Cover problem, where the goal is to compute a minimum set of vertices in a given graph which forms a vertex cover and induces a connected subgraph, is a fundamental combinatorial problem and has received extensive attention in various subdomains of algorithmics. In the area of kernelization, it is known that this problem is unlikely to have a polynomial kernelization algorithm. However, it has been shown in a recent work of Lokshtanov et al. (Proceedings of the 49th annual ACM SIGACT symposium on theory of computing, STOC 2017, Montreal, QC, Canada, 2017) that if one considered an appropriate notion of approximate kernelization, then this problem parameterized by the solution size does admit an approximate polynomial kernelization. In fact, Lokshtanov et al. were able to obtain a polynomial size approximate kernelization scheme (PSAKS) for Connected Vertex Cover parameterized by the solution size. A PSAKS is essentially a preprocessing algorithm whose error can be made arbitrarily close to 0. In this paper we revisit this problem, and consider parameters that are strictly smaller than the size of the solution and obtain the first polynomial size approximate kernelization schemes for the Connected Vertex Cover problem when parameterized by the deletion distance of the input graph to the class of cographs, the class of bounded treewidth graphs, and the class of all chordal graphs. [ABSTRACT FROM AUTHOR]

Floating Catchment Area (FCA) methods are a popular tool to investigate accessibility to public facilities, in particular health care services. FCA approaches are attractive because, unlike other accessibility measures, they take into account the potential for congestion of facilities. This is done by 1) considering the population within the catchment area of a facility to calculate a variable that measures level of service, and then 2) aggregating the level of service by population centers subject to catchment area constraints. In this paper we discuss an effect of FCA approaches, an artifact that we term demand and level of service inflation. These artifacts are present in previous implementations of FCA methods. We argue that inflation makes interpretation of estimates of accessibility difficult, which has possible deleterious consequences for decision making. Next, we propose a simple and intuitive approach to proportionally allocate demandand and level of service in FCA calculations. The approach is based on a standardization of the impedance matrix, similar to approaches popular in the spatial statistics and econometrics literature. The result is a more intiuitive measure of accessibility that 1) provides a local version of the provider-to-population ratio; and 2) preserves the level of demand and the level of supply in a system. We illustrate the relevant issues with some examples, and then empirically by means of a case study of accessibility to family physicians in the Hamilton Census Metropolitan Area (CMA), in Ontario, Canada. Results indicate that demand and supply inflation/deflation affect the interpretation of accessibility analysis using existing FCA methods, and that the proposed adjustment can lead to more intuitive results. [ABSTRACT FROM AUTHOR]

Freely-available satellite data streams and the ability to process these data on cloud-computing platforms such as Google Earth Engine have made frequent, large-scale landcover mapping at high resolution a real possibility. In this paper we apply these technologies, along with machine learning, to the mapping of peatlands–a landcover class that is critical for preserving biodiversity, helping to address climate change impacts, and providing ecosystem services, e.g., carbon storage–in the Boreal Forest Natural Region of Alberta, Canada. We outline a data-driven, scientific framework that: compiles large amounts of Earth observation data sets (radar, optical, and LiDAR); examines the extracted variables for suitability in peatland modelling; optimizes model parameterization; and finally, predicts peatland occurrence across a large boreal area (397, 958 km2) of Alberta at 10 m spatial resolution (equalling 3.9 billion pixels across Alberta). The resulting peatland occurrence model shows an accuracy of 87% and a kappa statistic of 0.57 when compared to our validation data set. Differentiating peatlands from mineral wetlands achieved an accuracy of 69% and kappa statistic of 0.37. This data-driven approach is applicable at large geopolitical scales (e.g., provincial, national) for wetland and landcover inventories that support long-term, responsible resource management. [ABSTRACT FROM AUTHOR]

Introduction: Quality of life is included in the economic evaluation of health services by measuring the preference for health states, i.e. health state utilities. However, most intervention studies include a disease-specific, not a utility, instrument. Consequently, there has been increasing use of statistical mapping algorithms which permit utilities to be estimated from a disease-specific instrument. The present paper provides such algorithms between the MacNew Heart Disease Quality of Life Questionnaire (MacNew) instrument and six multi-attribute utility (MAU) instruments, the Euroqol (EQ-5D), the Short Form 6D (SF-6D), the Health Utilities Index (HUI) 3, the Quality of Wellbeing (QWB), the 15D (15 Dimension) and the Assessment of Quality of Life (AQoL-8D). Methods: Heart disease patients and members of the healthy public were recruited from six countries. Non-parametric rank tests were used to compare subgroup utilities and MacNew scores. Mapping algorithms were estimated using three separate statistical techniques. Results: Mapping algorithms achieved a high degree of precision. Based on the mean absolute error and the intra class correlation the preferred mapping is MacNew into SF-6D or 15D. Using the R squared statistic the preferred mapping is MacNew into AQoL-8D. Implications for research: The algorithms reported in this paper enable MacNew data to be mapped into utilities predicted from any of six instruments. This permits studies which have included the MacNew to be used in cost utility analyses which, in turn, allows the comparison of services with interventions across the health system. [ABSTRACT FROM AUTHOR]

In this study, we used the Disaggregation based on Physical And Theoretical scale Change (DISPATCH) algorithm under very wet soil conditions in Western Canada for the disaggregation of coarse resolution 40-km soil moisture derived from the Soil Moisture Ocean Salinity (SMOS) satellite. The algorithm relies on the Soil Evaporative Efficiency (SEE), which was estimated using the 1-km resolution data from the MODerate resolution Imaging Spectoradiometer (MODIS). The study aimed to: (i) evaluate DISPATCH under wet soil conditions, (ii) test the linearity/non-linearity of the relationship between soil moisture and SEE, and (iii) propose a more robust procedure to calibrate the SEE model under very wet soil conditions. The disaggregated soil moisture values were compared to 0–5 cm in situ measurements and the soil moisture derived from the L-MEB (L-band Microwave Emission of the Biosphere) model from airborne brightness temperature at 1.4 GHz collected during the Canadian Experiment for Soil Moisture in 2010 (CanEx-SM10) field campaign. The results show a correlation between 0.7 and 0.8 and bias values ~ 0 m 3 /m 3 . The DISPATCH algorithm shows better disaggregation results under very wet soil conditions when a non-linear relationship is considered between SEE and soil moisture instead of a linear model. This is mainly due to the small variability of surface temperature inside the area covered by the SMOS pixel under very wet soil conditions, and the difficulty in accurately estimating the maximum soil temperature ( Ts max ), which is a driving factor for SEE. A sensitivity analysis was conducted and it shows that the linear model performs well only if Ts max can be determined more accurately. The possibility to determine Ts max using high resolution MODIS data over a larger area than the SMOS pixel is examined and discussed in the paper. [ABSTRACT FROM AUTHOR]

A heavy rainfall event over a 2-h period on 8 July 2013 caused significant flash flooding in the city of Toronto and produced 126 mm of rain accumulation at a gauge located near the Toronto Pearson International Airport. This paper evaluates the quantitative precipitation estimates from the nearby King City C-band dual-polarized radar (WKR). Horizontal reflectivity Z and differential reflectivity ZDR were corrected for attenuation using a modified ZPHI rain profiling algorithm, and rain rates R were calculated from R( Z) and R( Z, ZDR) algorithms. Specific differential phase KDP was used to compute rain rates from three R( KDP) algorithms, one modified to use positive and negative KDP, and an R( KDP, ZDR) algorithm. Additionally, specific attenuation at horizontal polarization A was used to calculate rates from the R( A) algorithm. High-temporal-resolution rain gauge data at 44 locations measured the surface rainfall every 5 min and produced total rainfall accumulations over the affected area. The nearby NEXRAD S-band dual-polarized radar at Buffalo, New York, provided rain-rate and storm accumulation estimates from R( Z) and S-band dual-polarimetric algorithm. These two datasets were used as references to evaluate the C-band estimates. Significant radome attenuation at WKR overshadowed the attenuation correction techniques and resulted in poor rainfall estimates from the R( Z) and R( Z, ZDR) algorithms. Rainfall estimation from the Brandes et al. R( KDP) and R( A) algorithms were superior to the other methods, and the derived storm total accumulation gave biases of 2.1 and −6.1 mm, respectively, with correlations of 0.94. The C-band estimates from the Brandes et al. R( KDP) and R( A) algorithms were comparable to the NEXRAD S-band estimates. [ABSTRACT FROM AUTHOR]

• A Survey is conducted on 693 EV owners in Canada pertaining to the EV usage. • Blockchain's smart contracts allow decentralized energy trading among EVs. • User preferences make the charging scheduling model adaptive to various conditions. • Peer-to-peer trading and ancillary services provision are increased by up to 90%. • Charging costs could be offset by external transactions revenue by up to 100%. Electric vehicles (EVs) are witnessing increased utilization throughout the world as an alternative to fossil–fueled vehicles. The extensive deployment of EVs can bring challenges to the grid if not properly integrated. Such challenges, however, could be exploited as opportunities if the huge unused capacity of the battery storage in millions of EVs are utilized for ancillary services to the grid and peer–to–peer (PtP) energy trade. Given that there is at least one human user per vehicle, human input must be considered to improve the scheduling process. To that end, this paper presents a new algorithm for bidirectional smart charging of EVs considering user preferences, PtP energy trade, and provision of ancillary services to the grid. The preferences of an EV user as input to the model are embedded into the scheduling process enabling the model to be adaptive to various conditions. Optimization slack variables are utilized for optimal management of EV battery SOC and energy allocation for multiple services. New indices are developed and introduced for quantification of the EV participation in ancillary services and PtP transactions. Real–world data has been collected and utilized for model specification and simulation to make the assumptions more realistic. The efficacy and feasibility of the proposed model are validated using numerical studies. The results indicate that incorporating users' preferences into the scheduling process would improve the aggregated revenue generated by the EV scheduling model which in turn could offset the charging costs by up to 100%. Further, an increase of about 90% in peer-to-peer energy transactions among EVs and 11% in ancillary services provision to the grid are achieved through the developed user-centric smart charging model. [ABSTRACT FROM AUTHOR]