The Variational Autoencoder (VAE) is a seminal approach in deep generative modeling with latent variables. Interpreting its reconstruction process as a nonlinear transformation of samples from the latent posterior distribution, we apply the Unscented Transform (UT) -- a well-known distribution approximation used in the Unscented Kalman Filter (UKF) from the field of filtering. A finite set of statistics called sigma points, sampled deterministically, provides a more informative and lower-variance posterior representation than the ubiquitous noise-scaling of the reparameterization trick, while ensuring higher-quality reconstruction. We further boost the performance by replacing the Kullback-Leibler (KL) divergence with the Wasserstein distribution metric that allows for a sharper posterior. Inspired by the two components, we derive a novel, deterministic-sampling flavor of the VAE, the Unscented Autoencoder (UAE), trained purely with regularization-like terms on the per-sample posterior. We empirically show competitive performance in Fr\'echet Inception Distance (FID) scores over closely-related models, in addition to a lower training variance than the VAE.

Uveal melanoma (UM) is the most common ocular malignancy in adults, associated with the poorest prognosis, with metastatic disease occurring in up to 50% of patients. In contrast to metastatic cutaneous melanoma, the use of immune checkpoint inhibitors is associated with poor outcomes in metastatic uveal melanoma (mUM). Tebentafusp, a bispecific molecule, has recently become the first treatment in decades to improve overall survival for mUM. This review summarises the existing and emerging immuno-oncology approaches for the treatment of mUM, and biomarkers of response and resistance to the same. Finally, we propose future research directions that could maximise treatment benefit to a wider pool of patients with UM.

BACKGROUND The efficacy of stromal vascular fraction (SVF) treatment, ie, stem cells, directly depends on the SVF cell count and their viability. The SVF cell count and viability are in direct correlation with adipose tissue harvesting site which yields SVF cells, thus making contribution to developing Tissue Guidance. OBJECTIVES To investigate the importance of harvesting subcutaneous adipose tissue-derived SVF cells on the concentration and viability of SVF. METHODS Adipose tissue was collected by vibration assisted liposuction, from the regions of the upper and lower abdomen, lumbar region and inner thigh region. Using the semi-automatic UNISTATION 2nd Version system (NeoGenesis, Seoul, South Korea), the obtained fat was chemically processed (collagenase enzyme) and a concentrate of SVF cells was obtained by centrifugation. These samples were then analyzed using the Luna-Stem Counter device (Logos Biosystems; Gyeonggi-do, South Korea) to measure the number and viability of SVF cells. RESULTS The highest concentration of SVF, comparing the regions of the upper abdomen, lower abdomen, lumbar, and inner thigh, was found in the lumbar region, specifically at 97498.00 per 1.0 ml of concentrate. The lowest concentration was found in the upper abdomen region. By ranking the viability values, the highest cell viability of SVF was observed in the lumbar region, measuring 36.6200%. The lowest viability was found in the upper abdomen region, measuring 24.4967%. CONCLUSIONS By comparing the upper and lower abdomen regions, lumbar and upper thigh regions, the authors have come to the conclusion that, on average, the largest number of cells and their viability was obtained from the lumbar region.

One of the most important challenges when building road infrastructure is the selection of appropriate mechanization, on which the efficiency of construction and the life of exploitation depends largely. As construction machinery, pavers occupy a significant place in civil engineering projects, so their selection, depending on a road category, is a very important activity. The objective of this paper is to develop an intelligent Fuzzy MCDM (Multi-Criteria Decision-Making) model, which consists of the integration of D and Z numbers for the selection of construction machinery. The IMF D-SWARA (Improved Fuzzy D Step-Wise Weight Assessment Ratio Analysis) method was used to determine weighting coefficients. A novel Fuzzy ARAS-Z (Additive Ratio Assessment) method has been developed to determine an adequate paver for a lower category of roads (asphalt width up to 5 m), which represents an important contribution and novelty of the paper. A total of 10 alternatives were evaluated based on 16 criteria which were classified into 4 main groups. The results have shown that the alternative A8—SUPER 1300-3 represents a paver with the best characteristics for the considered set of parameters. After that, verification tests were calculated, and they include a comparative analysis with four other MCDM methods based on Z numbers, a change in the normalization procedure, and the impact of changing the size of an initial fuzzy matrix. The tests showed the stability of the developed model with negligible deviations.

Over the last 14 years, ichthyological and ecological parameters have been monitored in the Labudovo okno Ramsar site. This area is important for its biodiversity as it is home to many rare and endangered plants and animal species. A total of 3861 fish specimens were sampled and measured at six sampling sites four times during the sampling period. An analysis of biodiversity indexes, relative biomass (kg/ha), and relative annual production (kg/ha) was carried out to assess the effectiveness of existing conservation measures. The results obtained show a trend decline in biodiversity, relative biomass, and relative annual production. This indicates a biodiversity conservation problem that should be addressed through other mechanisms in addition to the principles of the Ramsar Convention.

Different industries are observing the positive impact of 360 video on the user experience. However, the performance of VR systems continues to fall short of customer expectations. Therefore, more research into various design elements for VR streaming systems is required. This study introduces a SW tool that offers straight-forward encoding platforms to simplify the encoding of DASH VR videos. In addition, we developed a dataset composed of 9 VR videos encoded with seven tiling configurations, four segment durations, and up to four different bitrates. A corresponding tile size dataset is also provided, which can be utilised to power network simulations or trace-driven emulations. We analysed the traffic load of various films and encoding setups using the dataset that was presented. Our research indicates that, while smaller tile sizes reduce traffic load, video decoding may require more computational power.

Climate change is recognized as a global threat that negatively impacts biodiversity and forest resources. The use of existing indicators for sustainable forest management (SFM) related to biodiversity and climate change, as well as the development of new indicators, will help assess how forest management practices impact biodiversity enhancement and climate change mitigation. A Pan-European set of criteria and indicators has been developed as a policy instrument for monitoring, evaluating, and reporting on the progress in implementing SFM. In Bosnia and Herzegovina and Western Balkans in general, the Pan-European set of criteria and indicators is an insufficiently researched topic and there is a lack of scientific research conducted regarding their development and implementation. Through the analysis of the current situation in forestry of the Federation of Bosnia and Herzegovina (FBiH), regarding the compliance and importance of the Pan-European criteria for SFM, this paper aims to explain how the international process of development and application of the Pan-European criteria for SFM can contribute to the improvement of the situation in forestry and the creation of a consistent forest policy in FBiH. The survey among forestry professionals (n=360), from the public forest administration and public forest companies in FBiH, included the sets of questions related to socio-demographic characteristics, assessment of compliance and importance of six criteria of SFM. Research results revealed that forestry professionals are mainly males, on average 41 years old, with 13 years of working experience. The majority of forestry professionals in FBiH are not familiar with Pan-European criteria for SFM, and have a low level of their understanding. On average, forestry professionals indicated that the Pan-European criteria for SFM were of high importance, while compliance with current forest management activities were rated lower on average. The large differences between responses regarding the average rating of compliance and importance of the Pan-European criteria for SFM indicate their low level of implementation in FBiH forest management activities. Accordingly, the results indicate that there is a need to organize educational lifelong learning programs in FBiH forestry sector, involving forestry professionals and other interested parties, to generate knowledge related to the Pan-European criteria for SFM and the concept of SFM in general.

A novel deep operator network (DeepONet) with a residual U-Net (ResUNet) as the trunk network is devised to predict full-field highly nonlinear elastic-plastic stress response for complex geometries obtained from topology optimization under variable loads. The proposed DeepONet uses a ResUNet in the trunk to encode complex input geometries, and a fully-connected branch network encodes the parametric loads. Additional information fusion is introduced via an element-wise multiplication of the encoded latent space to improve prediction accuracy further. The performance of the proposed DeepONet was compared to two baseline models, a standalone ResUNet and a DeepONet with fully connected networks as the branch and trunk. The results show that ResUNet and the proposed DeepONet share comparable accuracy; both can predict the stress field and accurately identify stress concentration points. However, the novel DeepONet is more memory efficient and allows greater flexibility with framework architecture modifications. The DeepONet with fully connected networks suffers from high prediction error due to its inability to effectively encode the complex, varying geometry. Once trained, all three networks can predict the full stress distribution orders of magnitude faster than finite element simulations. The proposed network can quickly guide preliminary optimization, designs, sensitivity analysis, uncertainty quantification, and many other nonlinear analyses that require extensive forward evaluations with variable geometries, loads, and other parameters. This work marks the first time a ResUNet is used as the trunk network in the DeepONet architecture and the first time that DeepONet solves problems with complex, varying input geometries under parametric loads and elasto-plastic material behavior.

This paper provides a comprehensive overview of the latest Field Programmable Gate Array (FPGA) technologies that are being used to enhance smart and sustainable agriculture practices. The review was conducted using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) instructions as a report standard and different inclusion and exclusion criteria were utilized. To ensure data accuracy, reduce bias, and manage references, the Zotero tool was used. The agriculture industry is facing numerous challenges, such as water scarcity, soil degradation, and climate change, which can be mitigated through the integration of modern technologies. FPGAs have emerged as a promising solution in agriculture due to their high performance, flexibility, and low power consumption. In this paper, we discuss various FPGA-based solutions that are being developed to address key agricultural challenges, including precision farming, crop monitoring, livestock management, and environmental sensing. Authors also review FPGA architectures and programming tools that are specifically designed for agricultural applications. The use of FPGA-based systems in precision agriculture has resulted in numerous benefits, including reducing energy consumption, water usage, and temperature while expanding the range of applications. Additionally, the combination of IoT and FPGA is proving to be a promising solution for livestock management, and FPGA technology has also shown promise in energy efficiency and plant disease diagnosis. Overall, FPGA technology has proven to be a valuable tool in various aspects of agriculture, with its use expected to increase in the future.

Purpose This study aims to investigate the relationship between Islamic governance and the social performance of Islamic banks, pioneering a new aspect in terms of the impact of the National Shariah Board (NSB) on the social performance of Islamic banks. The essential body in the Islamic banks in charge of Islamic governance is the Shariah Supervisory Board (SSB). Therefore, in this study, the authors explore how the characteristics of the Shariah board and Islamic governance mechanisms influence the social performance of Islamic banks. Design/methodology/approach Panel data methods are applied to the annual data of 43 banks from 14 countries over the period 2012–2018 to explore the impact of Islamic governance on Islamic banks’ social performance. The authors have used all available bank annual reports in the given period. Social performance is measured by Maqasid al-Shariah (in terms of the goals of the Islamic moral economy) index using a comprehensive evaluation framework. Islamic governance is represented by the improved Islamic Governance Score (IG-Score) index, which measures the quality of Islamic governance in Islamic banks. In the research, the authors also introduce the frequency of SSB meetings in IG-Score. Findings The findings suggest a strong link between Islamic governance and the social performance of Islamic banks, illustrating the importance of the Shariah board in achieving maqasid. On the other hand, the research discovered that NSBs are inefficient and the existence of NSB can jeopardize the social performance of Islamic banks. The results of this research imply valuable recommendations for Islamic banks that are keen to improve their social performance. Originality/value Besides investigating the impact of SSB governance on the social performance of Islamic banks by using an improved IG score index, to the best of the authors’ knowledge, this is the first study that investigates the impact of NSBs on the social performance of Islamic banks.

Assessment of the functional significance of coronary artery stenosis using invasive measurement of fractional flow reserve (FFR) or non-hyperemic indices has been shown to be safe and effective in making clinical decisions on whether to perform percutaneous coronary intervention (PCI). Despite strong evidence from clinical trials, utilization of these techniques is still relatively low worldwide. This may be to some extent attributed to factors that are inherent to invasive measurements like prolongation of the procedure, side effects of drugs that induce hyperemia, additional steps that the operator should perform, the possibility to damage the vessel with the wire, and additional costs. During the last few years, there was a growing interest in the non-invasive assessment of coronary artery lesions, which may provide interventionalist with important physiological information regarding lesion severity and overcome some of the limitations. Several dedicated software solutions are available on the market that could provide an estimation of FFR using 3D reconstruction of the interrogated vessel derived from two separated angiographic projections taken during diagnostic coronary angiography. Furthermore, some of them use data about aortic pressure and frame count to more accurately calculate pressure drop (and FFR). The ideal non-invasive system should be integrated into the workflow of the cath lab and performed online (during the diagnostic procedure), thereby not prolonging procedural time significantly, and giving the operator additional information like vessel size, lesion length, and possible post-PCI FFR value. Following the development of these technologies, they were all evaluated in clinical trials where good correlation and agreement with invasive FFR (considered the gold standard) were demonstrated. Currently, only one trial (FAVOR III China) with clinical outcomes was completed and demonstrated that QFR-guided PCI may provide better results at 1-year follow-up as compared to the angiography-guided approach. We are awaiting the results of a few other trials with clinical outcomes that test the performance of these indices in guiding PCI against either FFR or angiography-based approach, in various clinical settings. Herein we will present an overview of the currently available data, a critical review of the major clinical trials, and further directions of development for the five most widely available non-invasive indices: QFR, vFFR, FFRangio, caFFR, and AccuFFRangio.

As the shipping sector has been one of the major impact factors on economic growth over the past decades, its digitalization is expected to make unprecedented improvements in the safety and reliability of ship control, thereby ultimately enabling the autonomous operations of ships. The automated control of ships will not only mitigate the risks of human mistakes but will also improve the efficiency of operations by preventing unexpected delays while being environmentally sustainable. With the advent of the Internet of Ships (IoS) sector, well-known and mature concepts of the Internet of Things (IoT) are being applied to ships and ports, thereby making them more and more equipped with sensing and communication capabilities that set the ground for improved situational awareness and better decision-making. However, there are many challenges that need to be thoroughly studied, such as the communication between barges, ports, and services, as increased network latency and limitations on the bandwidth imposed by satellite communications could introduce significant risks for accident occurrence, ultimately affecting the overall automated operation/teleoperation of barges. In this paper, we present one of the first attempts to test the potential of 5G systems for automating barge operations, starting from teleoperation as an enabler of automation, thereby creating and validating a cellular-based automated barge control system in a real-life environment. In this system, the barge is sailing in a busy port area such as one of the Port of Antwerp Bruges, while being connected to the 5G network. We assess the quality of the 5G communication system and present and discuss our initial results on the enhancements that 5G could bring to teleoperation and automation of the barge control.

5G Stand Alone (SA) networks are starting to be considered, designed and implemented in multiple countries in various forms (public, private, experimental). 5G SA networks mass adoption is expected to materialize by 2025. Mass deployment is anticipated at a large scale, due to the rich features and capabilities offered by 5G networks, including but not limited to slicing, service orchestration and automation, bringing the benefits of 5G among industry stakeholders and verticals. The concept of Network Applications is gaining momentum, as a way to ease the process of deploying industry-specific services and applications and to integrate them seamlessly with the new 5G networks and customer-specific application components. We target deploying and operating the novel 5G SA testbeds, Network Application and related capabilities in different T&L facilities across Europe. We envision the architectural advancement in terms of 5G features, such as orchestration, multi-slice implementation, Quality of Service (QoS)/Quality of Experience (QoE) and an innovative end-to-end monitoring framework, for network and application KPIs. In this paper, 5G open testbed advancements (3GPP Rel. 16 compliant) and readiness for Network Application experiments in real-life scenarios are presented, integrated as a unitary whole within the ED-funded VITAL-5G project.

The proliferation of 5G technology is enabling vertical industries to improve their day-to-day operations by leveraging enhanced Quality of Service (QoS). One of the key enablers for such 5G performance is network slicing, which allows telco operators to logically split the network into various virtualized networks, whose configuration and thus performance can be tailored to verticals and their low-latency and high throughput requirements. However, given the end-to-end perspective of 5G ecosystems where slicing needs to be applied on all network segments, including radio, edge, transport, and core, managing the deployment of slices is becoming excessively demanding. There are also various verticals with strict requirements that need to be fulfilled. Thus, in this paper, we focus on the solution for dynamic and quality-aware network slice management and orchestration, which is simultaneously orchestrating network slices that are deployed on top of the three 5G testbeds built for transport and logistics use cases. The slice orchestration system is dynamically interacting with the testbeds, while at the same time monitoring the real-time performance of allocated slices, which is triggering decisions to either allocate new slices or reconfigure the existing ones. In this paper, we illustrate the scenarios where dynamic provisioning of slices is required in one of the testbeds while taking into account specific latency/throughput/location requirements coming from the verticals and their end users.

This paper presents the implementation of the Binary Search Algorithm (BSA) to determine the Maximum Power Point (MPP) of a photovoltaic (PV) system under variable weather conditions. Additionally, the conventional well-known Perturb and Observe (P&O) algorithm is also implemented to be compared with the binary search based Maximum Power Point Tracking (MPPT) algorithm. Both algorithms are implemented in real time in MATLAB/Simulink environment. The experimental study is performed using the two 260 W series connected PV modules, the buck converter, and Humusoft MF 634 card to enable real-time operation. The value of the duty cycle for the buck converter is being updated in each step moving the operation point closer to MPP. The obtained experimental results demonstrate that the binary search based MPPT algorithm is more efficient and accurate when compared to the P&O MPPT algorithm.