Due to limited computational resources, acoustic models of early automatic speech recognition ( ASR ) systems were built in low-dimensional feature spaces that incur considerable information loss at the outset of the process. Several comparative studies of automatic and human speech recognition suggest that this information loss can adversely affect the robustness of ASR systems. To mitigate that and allow for learning of robust models, we propose a deep 2 D convolutional network in the waveform domain. The first layer of the network decomposes waveforms into frequency sub-bands, thereby representing them in a structured high-dimensional space. This is achieved by means of a parametric convolutional block defined via cosine modulations of compactly supported windows. The next layer embeds the waveform in an even higher-dimensional space of high-resolution spectro-temporal patterns, implemented via a 2 D convolutional block. This is followed by a gradual compression phase that selects most relevant spectro-temporal patterns using wide-pass 2 D filtering. Our results show that the approach significantly outperforms alternative waveform-based models on both noisy and spontaneous conversational speech ( 24% and 11% relative error reduction, respectively). Moreover, this study provides empirical evidence that learning directly from the waveform domain could be more effective than learning using hand-crafted features. by means of a non-parametric 2 D convolutional layer. is followed

From the 18th century, the Industrial Revolutions led to the creation of a base for the development and progress of civilization and business. Different business models that relied on the new trends defined by the industrial revolutions significantly transformed the ecosystems of companies. This resulted in the accelerated development and entrance into the digital age, followed by the digitization of the environment and establishing a new system of values. Information and communication technologies are the backbone of the development and mapping of the physical domain into cyber-physical, leading to the emergence of new processes, knowledge, and skills that have been established in the context of Digital Transformation and the digital era. The key part of Digital transformation process is the Smart Systems concept, which is present in almost all industries and domains. Smart Agriculture is one of the most common applications of Smart Systems solutions. It aims to answer important questions regarding food production while, at the same time, taking care of the preservation of the environment, the health of the population and global development. In this paper, the key concepts of the digital transformation and smart systems are presented. In addition, the design and development of the GoGrow, commercial solution for smart agriculture is described. Keywords-Industry 4.0; Digital Transformation; Smart Systems; Smart Agriculture;

Aim: Nosocomial infections (NIs) represent a major public health problem in developed, and even more in developing countries. Based on the origin of the pathogen, NIs are divided into endogenous caused by microorganisms of the patient’s microflora, and exogenous caused by pathogens from the patient’s environment or by the same microorganism isolated elsewhere from patient’s body. The main aim of this research was to determine the frequency, etiology and types of NIs at the Clinic for Surgery of the University Clinical Center Tuzla. Material and Methods: 5.039 patients were prospectively followed for the development of endogenous and exogenous NIs (January-December 2015). The definition of NI was performed using standardized the National Healthcare Safety Network (NHSN) criteria. Results: Based on continuous epidemiological surveillance, the incidence of NI was 3.51%; with a total of 177 registered infections, 24 endogenous and 153 exogenous. The most common NIs were urinary tract infections (UTIs) (14.29% endogenous and 85.71% exogenous) and surgical site infections (SSIs) (16.67% endogenous and 83.33% exogenous), p<0.001. Gram-negative bacteria were predominant (76.84%) over Gram-positive bacteria (23.16%). Gram-positive bacteria mainly caused bacteremia, while Gram-negative bacteria were most commonly isolated from UTIs and SSIs. The most common Gram-negative bacteria were Klebsiella pneumoniae (24.86%) and Pseudomonas aeruginosa (22.6%), and among Gram-positive, Staphylococcus aureus (10.73%) and coagulase-negative staphylococci (7.91%). Conclusion: Epidemiological surveillance is considered a key link in the program for the prevention and control of NIs. The most important, and the ultimate goal and purpose of conducting epidemiological surveillance are to reduce and eliminate the risk factors, which can lead to a reduction in NI incidence rate. Determining the endemic rates of NIs provides an objective understanding of the overall NI status in an institution as well as existing risk factors for the occurrence of these infections.