TorchEsegeta: Framework for Interpretability and Explainability of Image-based Deep Learning Models

Clinicians are often very sceptical about applying automatic image processing approaches, especially deep learning based methods, in practice. One main reason for this is the black-box nature of these approaches and the inherent problem of missing insights of the automatically derived decisions. In order to increase trust in these methods, this paper presents approaches that help to interpret and explain the results of deep learning algorithms by depicting the anatomical areas which influence the decision of the algorithm most. Moreover, this research presents a unified framework, TorchEsegeta, for applying various interpretability and explainability techniques for deep learning models and generate visual interpretations and explanations for clinicians to corroborate their clinical findings. In addition, this will aid in gaining confidence in such methods. The framework builds on existing interpretability and explainability techniques that are currently focusing on classification models, extending them to segmentation tasks. In addition, these methods have been adapted to 3D models for volumetric analysis. The proposed framework provides methods to quantitatively compare visual explanations using infidelity and sensitivity metrics. This framework can be used by data scientists to perform post-hoc interpretations and explanations of their models, develop more explainable tools and present the findings to clinicians to increase their faith in such models. The proposed framework was evaluated based on a use case scenario of vessel segmentation models trained on Time-of-fight (TOF) Magnetic Resonance Angiogram (MRA) images of the human brain. Quantitative and qualitative results of a comparative study of different models and interpretability methods are presented. Furthermore, this paper provides an extensive overview of several existing interpretability and explainability methods.

Jews, Blood, and Post-Zionist TV: The Mizrahi as Vampire in Juda

The success of the Israeli vampire–crime–comedy series Juda is not at all trivial, to say the least. It dared to adopt a controversial subgenre that is associated with antisemitism and blood libels. Moreover, it deals with social traumas and the ethnic conflict between the Zionist Ashkenazi hegemony and the Mizrahi sector, which accuses the hegemony of oppression and discrimination. Juda expresses a critical agenda: a dissolution of Zionist values as the only solution and chance for redemption, both for the hero and for society. Thus, despite emerging at a time when the horror genre had experienced a late blooming on Israeli screens, its appearance is connected to two other central processes in contemporary Israeli film and television: the incorporation of religion and the ascendancy of the Mizrahi hero. Juda overcomes the inherent problem in the image of the Jewish vampire—first by creating a distinction between a Jewish vampire and a gentile vampire, and second by having a protagonist who is a Mizrahi Jew.

Equine obesity and the role of the veterinary nurse

Equine obesity is defined as a medical disease in which excess body fat has accumulated to such an extent that it has an adverse effect on the general health of the horse. Obesity is a cause for concern, with one-third of the equine population in the UK being regarded as obese, although owner recognition of obesity in horses is an inherent problem, with many underestimating the body condition or weight of their horse. This is further complicated by the fact that with larger framed horses, or horses that are already overweight, assessing body condition is more difficult. There are a number of ways to assess body condition and the most practical means of regular assessment is body condition scoring, although this is regarded as subjective. As with many diseases and disorders, the cause of obesity is multifactorial. However, the most common reason for a horse to become obese is overfeeding, coupled with a lack of exercise. Obesity can be addressed with client education and veterinary nurses can provide advice on weight management programmes. However, these need to be tailored to the individual horse and owners need to recognise that they are entering into a long-term commitment.

‘Shuttle’ Technology for Noise Reduction and Efficiency Improvement of Hydrostatic Machines - Part 2

In 2001, INNAS introduced the ‘Shuttle’ technology for noise reduction and efficiency improvement of hydrostatic machines. The current study revisits this technology for application in hydrostatic pumps and motors. In many hydrostatic pumps and motors, commutation is imposed by a fixed component like a valve plate. Designing a valve plate (or comparable component) that ensures good commutation at one specific operating condition, is fairly simple. However, an inherent problem of such a component is that it should ensure good commutation at all of the operating conditions. In an attempt to minimise losses and reduce noise emission caused by improper commutation, so-called shuttles were introduced by INNAS in 2001. These shuttles act as small pistons between two working chambers, essentially providing a connection to the ports while the valve plate is still closed. In theory, this will result in a check-valve like commutation. In the original paper, shuttles were implemented in a hydraulic transformer. This paper discusses and analyses the use of shuttles in pumps and motors. Simulation results show that the introduction of shuttles can reduce commutation losses to negligible levels. Furthermore, the results suggest that the use of shuttles could also reduce noise emissions.

Analytical methods for the quantification of cisplatin, carboplatin, and oxaliplatin in various matrices over the last two decades

Background: Platinum derivatives including cisplatin and its later generations carboplatin, and oxaliplatin remain the most largely used drugs in the therapy of malignant diseases. They exert notable anticancer activity towards numerous types of solid tumors such as gastric, colorectal, bladder, ovary, and several others. The chemotherapeutic activity of these compounds, however, is associated with many unwanted side effects and drug resistance problems limiting their application and effectiveness. Proper dosage is still an inherent problem, as these drugs are usually prescribed in small doses. Objective: Several analytical methods have been reported for the accurate quantification of cisplatin, carboplatin, and oxaliplatin and their metabolites either alone or in combination with other chemotherapeutic drugs, in different matrices such as pharmaceutical formulations, biological fluids, cancer cells, and environmental samples. The main goal of this review is to systematically study the analytical methods already used for the analysis of cisplatin, carboplatin, and oxaliplatin in various matrices during the last two decades. Results and Conclusion: In the literature, reviews showed that numerous analytical methods such as electroanalytical, UV-visible spectrophotometry, chromatographic, fluorescence, atomic absorption spectrophotometry, and other spectroscopic methods combined with mass spectrometry were used for the determination of these compounds in various matrices.

VIKOR power law-based customer-oriented FMEA with complete unique risk priority numbers

PurposeAn inherent problem on risk priority number (RPN) value duplication of traditional failure modes and effect analysis (FMEA) also exists in two customer-oriented FMEAs. One has no unique value, and another has 1% unique values out of 4,000 possible values. The RPN value duplication has motivated the development of a new customer-oriented FMEA presented in this paper to achieve practically all 4,000 unique values and delivering reliable prioritization.Design/methodology/approachThe drastic improvement is the result of power-law and VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR). By having all three risk factors in a power-law form, all unique values can be obtained, and by applying VIKOR to these power-law terms, the prioritization is more practical and reliable.FindingsThe proposed VIKOR power law-based customer-oriented FMEA can achieve practically all 4,000 unique values and is tested with two case studies. The results are more logical than the results from the other two customer-oriented FMEAs.Research limitations/implicationsThe evaluation has been done on two case studies for the service sector. Therefore, additional case studies in other industrial sectors will be required to confirm the effectiveness of this new customer-oriented RPN calculation.Originality/valueAchieving all 1,000 unique values could only be done by having experts tabulate all possible combinations for the traditional FMEA. Therefore, achieving all 4,000 unique values will be much more challenging. A customer-oriented FMEA has been developed to achieve practically all 4,000 unique risk priority numbers, and that the prioritization is more practical and reliable. Furthermore, it has a connection to the traditional FMEA, which helps explain the traditional one from a broader perspective.

CIBERSECURITY ON INTERNET OF ROBOTICS THINGS: EXPERIMENTAL PLATFORM

The Internet of Robotic Things (IoRT) is a technology that looks for monitoring, operating, and maintaining the tasks of multiple robots through the cloud. However, using these robots in cyberspace has a risk and an inherent problem in cybersecurity. To analyze the implications of this technology, the objective was to design, operate and submit an IoRT system with the default configuration. The proposed methodology consisted of designing an IoRT architecture; implement three robotic platforms linked to the cloud, applying a sniffing and spoofing cyberattacks, assess the impacts, and propose solutions. The experiment used three prototypes: two servo motors, a 6-degree-of-freedom arm, and a workstation with a robot. Additionally, the tools of the experiment were a conventional computer, a Raspberry Pi microcomputer, the Robotic Operative System middleware, the Kali Linux distribution, and the ThingSpeak cloud service. The contributions of the work were three, first it was proven that four types of links are sufficient to homologate, and ensure the integrity, reliability, and availability in the operation of different types of robots. Also, it was possible the connection of these robots even though they are not designed to work on the internet through a slave-robot node link. Finally, a real list of the consequences was obtained, given the vulnerabilities and the attacks tested, as well as some recommendations.Keywords: Cybersecurity, IoRT, Industry 4.0., Common Vulnerabilities and Exposures, Cloud, ROS.

Symbolic and Numeric Kernel Division for GPU-based FEA Assembly of Regular Meshes with Modified Sparse Storage Formats

This paper presents an efficient strategy to perform the assembly stage of finite element analysis (FEA) on general-purpose graphics processing units (GPU). This strategy involves dividing the assembly task by using symbolic and numeric kernels, and thereby reducing the complexity of the standard single-kernel assembly approach. Two sparse storage formats based on the proposed strategy are also developed by modifying the existing sparse storage formats with the intention of removing the degrees of freedom-based redundancies in the global matrix. The inherent problem of race condition is resolved through the implementation of coloring and atomics. The proposed strategy is compared with the state-of-the-art GPU-based and CPU-based assembly techniques. These comparisons reveal a significant number of benefits in terms of reducing storage space requirements and execution time and increasing performance (GFLOPS). Moreover, using the proposed strategy, it is found that the coloring method is more effective compared to the atomics-based method for the existing as well as the modified storage formats.

Translating German philosophy into English: The case of Martin Heidegger

Language endangerment and language loss have become of focal interest for linguists and cultural anthropologists who bemoan the loss of linguistic diversity. The coinage of the term “linguicide” indicates the inherent problem that is related to mondialisation, universalization, and urbanization, which in itself is a highly controversial subject. The recent discoveries of Martin Heidegger’s black notebooks cast a new perspective on his work, revealing his revulsion at universalist ideologies and his antimodernism – and, most fatefully, his antisemitism: Jews who are to him the incarnation of rootlessness, distance from the soil, and thus subversion. Heidegger was born in a rural provincial German – and for many remained so, walking in the countryside, hating TV, airplanes, pop music, and processed food that all conspire to distract us from the basic wondrous nature of Being, overwhelming us with information, killing silence, and never leaving us alone, and thus keep us away from the confrontation with “das Nichts” (the Nothing), which lies on the other side of Being, that is, however, unknown to the chatter (das Gerede), which can be perceived in the newspapers, on TV and in the cities Heidegger hated to spend time in. Although he was a Nazi to the end, this does not mean that nothing can be learned from him or problems connected to his work. This library research deals with the complexity of translating this German philosopher into the English language. It draws not only on typical examples from Heidegger’s path-breaking philosophical work Sein und Zeit and presents attempts at translating it, but also points out their shortcomings and drawbacks. Additionally, it presents solutions to the problems that emerge from Heidegger’s idiosyncratic language. Generally speaking, it reveals the almost unbridgeable language barriers that can only be overcome at the expense of depth and authenticity. Homogenization can be seen as a way of leveling down ideas and concepts that end in language death.

A restaurant recommendation approach with the contextual information

Recommender systems can recommend products by analyzing the interests and habits of users. To make more efficient recommendation, the contextual information should be collected in recommendation algorithms. In the restaurant recommendation, the location and the current time of customers should also be considered to facilitate restaurants to find potential customers and give accurate and timely recommendations. However, the existing recommendation approaches often lack the consideration of the influence of time and location. Besides, the data sparsity is an inherent problem in the collaborative filtering algorithm. To address these problems, this paper proposes a recommendation approach which combines the contextual information including time, price and location. Instead of constructing the user-restaurant scoring matrix, the proposed approach clusters price tags and generates the user-price scoring matrix to alleviate the sparsity of data. The experiment on Foursquare dataset shows that the proposed approach has a better performance than traditional ones.