In this paper, we consider the problem of sparsifying BERT models, which are a key building block for natural language processing, in order to reduce their storage and computational cost. We introduce the Optimal BERT Surgeon (oBERT), an efficient and accurate pruning method based on approximate second-order information, which we show to yield state-of-the-art results in both stages of language tasks: pre-training and fine-tuning. Specifically, oBERT extends existing work on second-order pruning by allowing for pruning weight blocks, and is the first such method that is applicable at BERT scale. Second, we investigate compounding compression approaches to obtain highly compressed but accurate models for deployment on edge devices. These models significantly push boundaries of the current state-of-the-art sparse BERT models with respect to all metrics: model size, inference speed and task accuracy. For example, relative to the dense BERT-base, we obtain 10x model size compression with < 1% accuracy drop, 10x CPU-inference speedup with < 2% accuracy drop, and 29x CPU-inference speedup with < 7.5% accuracy drop. Our code, fully integrated with Transformers and SparseML, is available at https://github.com/neuralmagic/sparseml/tree/main/research/optimal_BERT_surgeon_oBERT.

The progress in developing quantum hardware with functional quantum processors integrating tens of noisy qubits, together with the availability of near-term quantum algorithms has led to the release of the first quantum computers. These quantum computing systems already integrate different software and hardware components of the so-called “full-stack”, bridging quantum applications to quantum devices. In this paper, we will provide an overview on current full-stack quantum computing systems. We will emphasize the need for tight co-design among adjacent layers as well as vertical cross-layer design to extract the most from noisy intermediate-scale quantum (NISQ) processors which are both error-prone and severely constrained in resources. As an example of co-design, we will focus on the development of hardware-aware and algorithm-driven compilation techniques.

Abstract Our aim was to explore and summarize available cases of Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) suspected to be associated with amoxicillin reported in the literature. Electronic searches were conducted in several databases. Fifty-one publications describing a total of 64 patients who satisfied inclusion criteria were included in the review. The age of the patients ranged from 1.5-80 years (median: 24.5 years). TEN, SJS and SJS/TEN overlap were diagnosed in 30 (46.9%), 28 (43.8%) and 1 (1.6%) patients, respectively. SJS/TEN may occur promptly after administration of amoxicillin, but it could also be a delayed adverse effect. The total length of hospital stay ranged from 3-70 days (median: 16 days). Amoxicillin-induced SJS/TEN is accompanied by frequent occurrence of serious complications, long-term ocular and skin sequelae and high mortality rate. Clinicians should be aware that amoxicillin alone or combined with clavulanic acid can cause SJS/TEN in patients of all ages.

ABSTRACT The aim of the present systematic review was to critically analyse the relationship between tumour suppressor genes (TSGs) promoter methylation, a potent mechanism of gene silencing, and the development of salivary gland tumours, as well as the possible effect on clinical/histological characteristics. Review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (registration ID CRD42020218511). A comprehensive search of Web of Science, Scopus, PubMed, and Cochrane Central Register of Controlled Trials was performed utilizing relevant key terms, supplemented by a search of grey literature. Newcastle-Ottawa Quality Assessment Scale (NOQAS) was used for the quality assessment of included studies. Sixteen cross-sectional and 12 case-control studies were included in the review, predominantly dealing with methylation in TSGs related to DNA repair, cell cycle, and cell growth regulation and differentiation. Quantitative synthesis could be performed on P16 (inhibitor of cyclin-dependent kinase 4a), RASSF1A (Ras association domain family 1 isoform A) and MGMT (O6-methylguanine DNA methyltransferase) genes only. It showed that P16 and RASSF1A genes were more frequently methylated in salivary gland tumours compared to controls (P = .0002 and P < .0001, respectively), while no significant difference was observed for MGMT. Additionally, P16 did not appear to be related to malignant transformation of pleomorphic adenomas (P = .330). In conclusion, TSG methylation is involved in salivary gland tumour pathogenesis and several genes might play a considerable role. Further studies are needed for a better understanding of complex epigenetic deregulation during salivary gland tumour development and progression.

The vertex (respectively edge) metric dimension of a graph G is the size of a smallest vertex set in G, which distinguishes all pairs of vertices (respectively edges) in G, and it is denoted by dim(G) (respectively edim(G)). The upper bounds dim(G)≤2c(G)−1 and edim(G)≤2c(G)−1, where c(G) denotes the cyclomatic number of G, were established to hold for cacti without leaves distinct from cycles, and moreover, all leafless cacti that attain the bounds were characterized. It was further conjectured that the same bounds hold for general connected graphs without leaves, and this conjecture was supported by showing that the problem reduces to 2-connected graphs. In this paper, we focus on Θ-graphs, as the most simple 2-connected graphs distinct from the cycle, and show that the the upper bound 2c(G)−1 holds for both metric dimensions of Θ-graphs; we characterize all Θ-graphs for which the bound is attained. We conclude by conjecturing that there are no other extremal graphs for the bound 2c(G)−1 in the class of leafless graphs besides already known extremal cacti and extremal Θ-graphs mentioned here.

Objective: Herein we report an unusual case of torsed giant cystic Meckel’s diverticulum (MD) in a three-year-old boy, illustrating the rarity of this condition and briefly reviewing the literature on it. Case Report: A three-year-old boy who clinically presented with severe colicky abdominal pain and frequent vomiting. Abdominal ultrasonography showed a cystic lesion in the lower abdomen. Contrast-enhanced CT revealed a well-defined right upper quadrant cystic mass measuring ~5 × 4 cm. The cystic mass appeared to be arising from the small intestine. The radiological features were not suggestive of malignancy. An urgent laparotomy was performed. At operation, he was found to have an incomplete torsed giant cystic MD. Postoperative recovery was uneventful, and the patient was discharged on the sixth postoperative day. A regular, 3-year follow-up revealed no complications. Conclusion: A giant cystic Meckel’s diverticulum represents a rare cause of intra-abdominal/pelvic mass and should be considered in its differential diagnosis. Timely surgical intervention reduces morbidity and provides the best

Growing insecurity in contemporary society has helped make lifelong learning a prominent feature of postmodern discourses on education and society. As an independent discourse, lifelong learning has generated diversified learning contexts and a proliferation of education provision, issuing a serious challenge to traditional formal and institutional models of education. As a result, higher education, long enclosed in an ivory tower of elitist scholarship and accessible to only a few privileged individuals, has started to change course and open up to non-traditional disciplines, learners, and students. The trend to massification has pushed it to embrace students of various backgrounds and levels of academic preparedness. Policies have been created that facilitate implementation of higher education’s “lifelong dimension”, just as models have been designed that fit programs ranging from vocational to liberaleducation. This paper applies this framework to examine models for incorporating lifelong learning provision into Islamic education at tertiary level. Concluding remarks stress the significance of lifelong learning opportunities in Islamic education, especially given current demand for higher education as a way to meet the diverse needs of society and the public promptly.

Real-life mobility, also called “enacted” mobility, characterizes an individual’s activity and participation in the community. Real-life mobility may be facilitated or hindered by a variety of factors, such as physical abilities, cognitive function, psychosocial aspects, and external environment characteristics. Advances in technology have allowed for objective quantification of real-life mobility using wearable sensors, specifically, accelerometry and global positioning systems (GPSs). In this review article, first, we summarize the common mobility measures extracted from accelerometry and GPS. Second, we summarize studies assessing the associations of facilitators and barriers influencing mobility of community-dwelling older adults with mobility measures from sensor technology. We found the most used accelerometry measures focus on the duration and intensity of activity in daily life. Gait quality measures, e.g., cadence, variability, and symmetry, are not usually included. GPS has been used to investigate mobility behavior, such as spatial and temporal measures of path traveled, location nodes traversed, and mode of transportation. Factors of note that facilitate/hinder community mobility were cognition and psychosocial influences. Fewer studies have included the influence of external environments, such as sidewalk quality, and socio-economic status in defining enacted mobility. Increasing our understanding of the facilitators and barriers to enacted mobility can inform wearable technology-enabled interventions targeted at delaying mobility-related disability and improving participation of older adults in the community.

Taha Džabir el-Alvani, Reviving the Balance: The Authority of the Qurʼan and the Status of the Sunnah. Herndon: Internacionalni institut za islamsku misao, 1438/2017. Str. 210. ISBN 978-1565646902.

In a data‐driven economy, virtually all industries benefit from advances in information technology—powerful computing systems are critically important for rapid technological progress. However, this progress might be at risk of slowing down if the discrepancy between the current computing power demands and what the existing technologies can offer is not addressed. Key limitations to improving energy efficiency are the excessive growth of data transfer costs associated with the von Neumann architecture and the fundamental limits of complementary metal–oxide–semiconductor (CMOS) technologies, such as transistors. Herein, three approaches that will likely play an essential role in future computing systems are discussed: memristive electronics, spintronics, and electronics based on 2D materials. The authors present how these technologies may transform conventional digital computers and contribute to the adoption of new paradigms, like neuromorphic computing.

Anne Norton, On the Muslim Question. Princeton: Princeton University Press, 2013. Str. 272.

Abstract Diversity management in the workplace is essential for ensuring a high quality, representative public service and strengthening organizations' sustainability. For the public sector to remain effective and productive, the management of today’s dynamic workforce requires a perspective that values and welcomes diversity. The promotion of diversity in the public sector helps to ensure sustained competitiveness and success. This study explores managing diversity in the public sector and how public organizations' human resources have engaged with diversity. Semi-structured interviews with 37 officials working in Human Resource departments in the public sector, primarily in Northwest Florida, USA, were conducted by student researchers. A grounded theory approach to understanding the interviews was conducted, using two-cycle coding. Interviewees indicated that diversity recruitment and diversity education ensured high inclusion levels concerning race/ethnicity, age, and gender in the public sector, echoing the literature on the topic; this showed the importance of maintaining such emphases. Analysis of the interviews indicated continuing challenges with attaining the ideals of diversity. By providing attention to workforce diversity and insights into diversity management in the public sector organizations in the United States deserving of additional attention, this study contributes to understanding managing diversity to attain public sector success.

Abstract This study analyses the relationship of environmental sustainability and the corporate social responsibility (CSR) of business schools by using the partial least squares structural equation modelling (PLS-SEM) empirical approach on a sample of 338 students from South East Europe. In support of the extant theory of responsible management education, emphasizing the transdisciplinary relationship between the Ethics, CSR, and Sustainability (ERS) domains, we found a direct relationship between environmental sustainability and CSR of business schools. However, we empirically verified a path of indirect effects at the institutional level, starting with the idealism of individual students, leading to the CSR institutional involvement of a business school, mediated by its environmental involvement. Provided that the idealistic individuals might be driving the functioning of the individual responsible management education and its domains, we propose the existence of a potential halo effect ('ERS halo effect'), which has already been described and verified in the corporate sector. We believe that its dynamics, based on the biased assessment of a single business school ERS domain, with its outcomes reflected in the other domains, should be further explored in different institutional and cultural environments.

Patients with cancer have a higher risk of severe coronavirus disease (COVID-19) and associated mortality than the general population. Owing to this increased risk, patients with cancer have been prioritized for COVID-19 vaccination globally, for both primary and booster vaccinations. However, given that these patients were not included in the pivotal clinical trials, considerable uncertainty remains regarding vaccine efficacy, and the extent of humoral and cellular immune responses in these patients, as well as the risks of vaccine-related adverse events. In this Review, we summarize the current knowledge generated in studies conducted since COVID-19 vaccines first became available. We also highlight critical points that might affect vaccine efficacy in patients with cancer in the future. Vaccination against COVID-19 confers robust protection from severe disease. However, the extent to which this applies to patients with cancer remains uncertain given that these patients were excluded from most of the pivotal studies. In this Review, the authors provide an overview of the efficacy and immunogenicity of COVID-19 vaccines in patients with cancer, and discuss alternatives to vaccination for those who might be unable to develop a proficient immune response following vaccination. Vaccination against COVID-19 administered according to current prime–boost concepts is both safe and clinically effective in patients with cancer. To date, no reliable correlate of protection that allows the definite deduction of clinical efficacy from immune responses has been established, either in patients with cancer or in the general population. Patient-associated factors such as advanced age, haematological malignancy and/or treatment-associated factors such as B cell depletion might all lead to less proficient immune responses following vaccination. Future research will determine the necessity of further booster regimens as well as therapeutic options for those who do not benefit from active COVID-19 vaccination. Vaccination against COVID-19 administered according to current prime–boost concepts is both safe and clinically effective in patients with cancer. To date, no reliable correlate of protection that allows the definite deduction of clinical efficacy from immune responses has been established, either in patients with cancer or in the general population. Patient-associated factors such as advanced age, haematological malignancy and/or treatment-associated factors such as B cell depletion might all lead to less proficient immune responses following vaccination. Future research will determine the necessity of further booster regimens as well as therapeutic options for those who do not benefit from active COVID-19 vaccination.