Resistive Random Access Memory (RRAM) is a promising technology for power efficient hardware in applications of artificial intelligence (AI) and machine learning (ML) implemented in non-von Neumann architectures. However, there is an unanswered question if the device non-idealities preclude the use of RRAM devices in this potentially disruptive technology. Here we investigate the question for the case of inference. Using experimental results from silicon oxide (SiOx) RRAM devices, that we use as proxies for physical weights, we demonstrate that acceptable accuracies in classification of handwritten digits (MNIST data set) can be achieved using non-ideal devices. We find that, for this test, the ratio of the high- and low-resistance device states is a crucial determinant of classification accuracy, with ~96.8% accuracy achievable for ratios >3, compared to ~97.3% accuracy achieved with ideal weights. Further, we investigate the effects of a finite number of discrete resistance states, sub-100% device yield, devices stuck at one of the resistance states, current/voltage non-linearities, programming non-linearities and device-to-device variability. Detailed analysis of the effects of the non-idealities will better inform the need for the optimization of particular device properties.

In a continuous system of production in open-cast coal mining, in addition to the basic machinery which consists of highly productive machines (bucket-wheel excavators, bucket-chain excavators, spreaders, conveyor-belt transporters, etc.) that provide an uninterrupted exploitation of mineral ores, the usage of auxiliary machinery is also necessary. The auxiliary machinery provides the possibility of using the capacities of basic machinery adequately and timely. Further, it must be mentioned that the auxiliary machinery generates up to 20% of the total mining expenses. Dozers are the most commonly used machines. Choosing the adequate dozer of the right type and manufacturer is of great importance and can be made easier by carrying out an analysis of the existent operational-exploitative parameters. In this paper, the method of grading the operational-exploitative parameters of dozer is defined. The choice of the optimal type of dozer was conducted by using the analytical hierarchy process (AHP).

For many years, the computer industry has relied on steady progress in the exponential rate of scaling MOSFETs in integrated circuits. The usual expectation, based on Moore's law, is that the number of transistors able to be packed on a chip doubles roughly every 18 months. Sustaining this pace requires aggressive research into the numerous bottlenecks that threaten to slow it down. Much research has gone into the photolithography needed to produce such dense circuits, device structures that would allow smaller channel lengths, and a plethora of other materials and device advances that help sustain the present rate of scaling. In the past decade, however, another issue has emerged that threatens to impose an absolute limit on how many transistors can be packed onto a die. This is the issue of heat dissipation.

Summary Background: During reproductive life of women, adenosine causes both contraction (with low concentrations) of fallopian tubes and inhibition of their spontaneous motor activity (with high concentrations). Objective: The aim of this study was to investigate effects of natural agonists of adenosine, P2X and P2Y receptors on motility of isolated fallopian tubes taken from postmenopausal women. Materials and Methods: Isolated preparations of isthmus and ampoule were made from fallopian tubes of 21 women in post-menopause, and then tested for reactivity on increasing concentrations of adenosine and P2X/P2Y selective agonists. Results: Adenosine showed concentration-dependent inhibitory effect on spontaneous contraction of both isthmic and ampullary segments of fallopian tubes, while P2X and P2Y agonists (adenosine-5-diphosphate, adenosine-5-triphosphate, uridine-5-diphosphate, and uridine-5-triphosphate) did not influence motility of the isolated preparations. Contractile effect of adenosine was not observed throughout the concentration range used in the experiments. Conclusions: Fallopian tubes of postmenopausal women are unresponsive to P2X and P2Y agonists, unlike those of women in reproductive period. Only an inhibitory effect of adenosine on spontaneous contractions of fallopian tubes is maintained in post-menopause, while a contractile effect is observed in younger women at low concentrations is lost.

Fiber tracking with diffusion‐weighted MRI has become an essential tool for estimating in vivo brain white matter architecture. Fiber tracking results are sensitive to the choice of processing method and tracking criteria.

One of the basic and strategic goals of post-war Bosnia and Herzegovina is a full membership in the European Union. During 1992-1995, Bosnia and Herzegovina has passed through one of the most difficult periods in its’ history. The war has completely devastated the economical structure and the aftermaths were visible in the following years. As a new paradigm, the European Union appears. The European Union has helped the Western Balkans countries to overcome national tensions and offered them a new perspective; new regional and european integrations should be a new, historical chance for them. In the light of this, the focus of this paper is the impact of trade liberalization between the European Union and Bosnia and Herzegovina on the export from Bosnia and Herzegovina to European Union. The above mentioned impact has been observed in the period of eleven years. The total export amount of Bosnia and Herzegovina in 2015 was 4,6 billion euros, while the amount of the import was 8,1 billion euros, which makes nearly 60% in favor of import. Traditionally, Bosnia and Herzegovina’s most important foreign trade partner is the European Union where Bosnia and Herzegovina has a deficit in foreign trade. The portion of export to EU during 2015 was 71,79% while the portion of import was 60,85%. The coverage of import over export was 66,89%. In the period of 2005-2015, most of the products have been placed to markets in Croatia, Germany, Italy, and Serbia, which is not the EU member. The process of trade liberalization in 2015 has resulted in surplus of trade with certain EU countries; Austria, Luxembourg and Estonia while the decrease in trade has been recorded with Germany, Slovenia and Czech Republic. Regarding the export, Bosnia and Herzegovina mailny places products which have labour-intensive character. In the econometrics analysis, extented gravitational models of trade have been placed. Besides the abovementioned, the following econometrical techinques have been used: panel data models (PD), fixed effecs models (RE) and panel corrected standard errors (PCSE). In addition, Hausman test has been applied. According to the analysis that has been carried out, the following has been concluded; InGDPi, InGDPj and InDISTij variables are statistically important on the level of 5% in the period of gradual liberalization. The SSP1 an EK variables are not siginificant.In the other moder, in the period of complete liberalization InGDPi, InGDPj and SSP2 variables have a significant impact on a dependent variable on a level of 5%. InDISTij and EK variables have a significant impact on a level od 5% as well. The above mentioned result confirm the research hypothesis; trade liberalization has a significant impact on the export of Bosnia and Herzegovina to the EU. The basic limit of this research can be a relatively short observation period, caused by the fact that the data of merchandise trade between Bosnia and Herzegovina and EU countries until 2005 are not accessible.

Planning and Learning are complementary approaches. Planning relies on deliberative reasoning about the current state and sequence of future reachable states to solve the problem. Learning, on the other hand, is focused on improving system performance based on experience or available data. Learning to improve the performance of planning based on experience in similar, previously solved problems, is ongoing research. One approach is to learn Value function (cost-to-go) which can be used as heuristics for speeding up search-based planning. Existing approaches in this direction use the results of the previous search for learning the heuristics. In this work, we present a search-inspired approach of systematic model exploration for the learning of the value function which does not stop when a plan is available but rather prolongs search such that not only resulting optimal path is used but also extended region around the optimal path. This, in turn, improves both the efficiency and robustness of successive planning. Additionally, the effect of losing admissibility by using ML heuristic is managed by bounding ML with other admissible heuristics.

In the turbulent environment which is characterized by constant market changes and the development of informational technologies, supply chain is becoming the key instrument for competitive advantage. To survive on the market and be competitive, enterprises have to share and distribute knowledge and information. Information sharing affects the fundamental decisions of the supply chain management. Information sharing with partners is a precondition for the exchange of knowledge needed for business. Most important partners for every enterprise are their suppliers and buyers. Suppliers and buyers are key participants of the supply chain from which they get needed information. Relations with suppliers and buyers are an important precondition for the improvement of the operative performance of the enterprise. Sharing information with suppliers and buyers is done in the following way: information delivered to suppliers, information obtained from the supplier, information delivered to buyers and information obtained from buyers. To examine the impact of the supply chain used an empirical study about multi-dimensional relations. The research is focused on information sharing by food industry in Bosnia and Herzegovina (BiH). A random systematic sample was used to distribute the questionnaire to these companies. Collected data is analysed using the Confirmatory factor analysis (CFA), and the model was tested using the Structural equations model (SEM). The results showed that information sharing impacts the improvement of the organizational performance of companies, and partnerships with supplier and buyer impact on the information sharing. Results of this study showed that information sharing does not have a significant impact on the sub constructors of operative performances of the enterprise because there is no significant connection with flexibility, while for the other constructors there is a significant connection between information sharing and operative performances of the enterprise. Based on these facts, it examined the importance information sharing in the supply chain for the development of partnerships and improvement of performance this companies. The results will help managers on food industry in BiH how to improve quality of the information sharing through partnerships and how to developing operative performances companies. The model gives directions for developing business enterprises in food industry using information sharing within the supply chain. Apart from that, obtained results contributed to the better understanding of the significance of the information sharing in the food supply industry enterprises. Based on what is said, to improve the business of the enterprise, it is needed to share quality information which is possible to get through improving cooperation with key participants in the supply chain.