Several recent works show impressive results in mapping language-based human commands and image scene observations to direct robot executable policies (e.g., pick and place poses). However, these approaches do not consider the uncertainty of the trained policy and simply always execute actions that are suggested by the current policy as the most probable ones. This makes them vulnerable to domain shift and inefficient in the number of required demonstrations. We extend previous works and present the PARTNR algorithm that can detect ambiguities in the trained policy by analyzing multiple modes in the probability distributio of pick and place poses using topological analysis. In this way uncertainty in action can be estimated with single inference (and training single model) instead of using ensemble of models. Additionally, PARTNR employs an adaptive, sensitivity-based, gating function that decides if additional user demonstrations are required. User demonstrations are aggregated to the dataset and used for subsequent training. In this way, the policy can adapt promptly to domain shift and it can minimize the number of required demonstrations for a well-trained policy. The adaptive threshold enables to achieve the user-acceptable level of ambiguity to execute the policy autonomously and in turn, increase the trustworthiness of our system. We demonstrate the performance of PARTNR in a table-top pick and place task.

Abstract Innovation is crucial for the competitiveness of countries in the global market. Countries oriented to progress must invest in innovative activities. Using the example of the Western Balkan countries, this study investigated their innovative competitiveness. The indicators of the Global Innovation Index (GII) were used for the years 2019, 2020, and 2021. A multi-criteria approach was based on the innovative SMART approach. The weights of the criteria were determined using the CRITIC (criteria importance through intercriteria correlation) method, while the CRADIS (compromise ranking of alternatives from distance to ideal solution) method ranked the Western Balkan countries according to their innovative characteristics. The results showed that Montenegro had the best innovation indicators, followed by Serbia, while Albania had the worst indicators. This research showed an innovative approach of using four normalizations, and the advantages of these normalizations were used to contribute to stabilization in multi-criteria decision-making.

Flight to the Moon (1969), according to many researchers of the history of science and civilization, is the greatest scientific and technological achievement of the human race. Thus, a centuries-old dream came true: man left Earth for the first time and stepped on another celestial body. Stunned by the achieved success, the scientists set themselves new, more difficult goals, which seemed feasible to them: conquering the planets of the solar system and migrating life to them. And even more and even harder: going beyond the solar system, deep into Cosmos. Analyzing exactly this, undoubtedly, extraordinary success of human thought, and extending it to the scale of the Cosmos, we conclude in this paper that it is a “false symbol”: The moon is so close to Earth that man's flight to the moon is not a flight to Cosmos. We have shown by calculation that, due to the constants that God has determined perfectly precisely and according to which the Cosmos functions (and which, regardless of the development of science, we cannot change), man will never travel through the Cosmos! Thus, the circle is closed: science has led us to God’s instructions as eternal truths given in the Holy Books: God ordained the earth for man’s life.

Southern Europe is an exceptionally fire-prone region. Effective prevention, early warning, and early response are important actions that are needed to reduce disastrous consequences on people and structures. Flame pixel classification is the first and often critically important stage of fire detection systems based on computer vision. In this paper, a number of rule-based flame pixel classifiers in still images are studied based on the BoWFire fire dataset. The performances of the considered rule-based flame pixel classifiers are reported in terms of F1 score, Matthews correlation coefficient, and Balanced accuracy. The experimental results demonstrate the effectiveness of simple rule-based flame pixel classifiers for flame detection in still images.

Phenological research of plant species is of great importance in the context of adaptation to climate change and changing environmental factors, especially in dynamic urban environments, such as the area of Sarajevo. This research aims to determine trends in the phenological pattern of hybrid plane trees in the area of Sarajevo so that recommendations can be made for the use of plane trees in greening urban and suburban areas since they largely depend on microclimatic conditions. In this paper, the authors researched the variability of leafing phenology of maple (Platanus × acerifolia (Ait) (Wild)) at six different localities in the area of Sarajevo. Observations were made in the spring of 2009, 2014, 2016, and 2020. Six phenological phases in the spring aspect of leaf development were monitored (0 - dormant buds, 1 - beginning of bud opening, 2 - open buds, 3 - leaf opening, 4 - young leaves, 5 - fully developed leaves). The results showed differences in the beginning and end of phenological phases by years and localities. Analysis of variance showed statistically significant differences in the duration of leaf development phases caused by the year of observation, locality, and the interaction of locality and year, which indicates the influence of seasonal climatic elements and micro-location conditions, as well as their interaction on the occurrence of phenophases. The results of this research can be used to recommend the use of plane trees in selected locations, with the selection of appropriate provenances and respect for phenological characteristics. Research needs to be continued and extended to leaf rejection research, which is particularly significant given the frequent heavy snowfall during the winter months in the investigated area.

This paper describes the pilot implementation of blockchain technology (Ethereum) for smart grid data management in the IT environment of electrical distribution company Elektro Celje. The work focuses on the decentralised notation of smart meter data to enable secure access and prevent data misuse. The procedure for setting up the pilot project and the operational functions as well as the results of the performance tests are presented and discussed.

The paper presents an early prototype of an online tool named SpeeD, which is aimed at conversion of a recorded speech into the corresponding conceptual database model. SpeeD is the first tool that provides such functionality. The tool is currently able to process English speech and generate the model represented by the corresponding UML class diagram.

The study of Bosniak literature in the period of the Independent State of Croatia has been marginalized in previous literary-historical studies, and the reasons for this were ideological and political in nature, and not scientific. This work deals with the status of Bosniak literature in the literary-critical horizons of Vladimir Jurčić, the bellwether of the Ustasha national ideology in Bosnia and Sarajevo, in the period from 1941st to 1945th. As a professor, editor of daily and periodical publications, he wrote about Bosniak literature and its canonical writers in the light of the ideological and political worldviews. He propagated theses about socio-political function of literature that extends „people's spirit”, „racial-biological” and „national” features. Jurčić attributed to literature a mediating role in transmitting the deep identity of the Croatian people, and developing a thesis on the Croatian national identity of Bosnian Muslims (Bosniaks) he treated Bosniak writers as the most representative reflectors of Croatian national consciousness in Bosnia. In addition to individual studies on Bosniak writers, Jurčić stated that they were separate units of the unpublished book Muslims in Croatian Literature. Jurčić's literary critical habitus is a product of socio-political and intellectual circumstances in Croatia - in the narrower sense and in the SHS - in the broader sense, which were used as a starting point for the production of certain ideological, political and cultural values in the NDH. As a follower of the ideological platform of Radić's HSS (peasant movement) and its reflections on discursive practices, especially in the social - humanities sciences (Dvorniković, Radić, Tomašić, Lukas), he interpreted literature in accordance with these practices, reducing its meaning only to ruling ideologues. He valorized Bosniak literature as a component of Croatian literature, applying several criteria: collective, linguistic, territorial and religious, which he sought to include the widest possible range of identity features and thus support the thesis of Croatianness Bosnian Muslims (Bosniaks). In literary criticism, he promoted theses on racial, ethical and eugenic superiority, then on the national spirit, linguistic and stylistic specifics of Bosnian Muslims (Bosniaks) as an „organic“ part of the Croatian people. He emphasized the „poljodjelski“ character of Bosniak writers between the two world wars, while in older literature, especially in the oral literary tradition- and all that for need of ideological manipulation in the time of the Independent State of Croatia - war, he emphasized the highland (tribal) character that manifested itself in the epic-agonal consciousness. All these theses arose from the idea of unity and continuity of the „organic nation“, but did not find a stronghold in Bosnia because it was cultural and historical terms different from the native Croatian space, which was in principle a fundamental obstacle to its realization. Aware of the insurmountability of the cultural, literary and historical uniqueness of Bosnia, Jurčić constructed and established the literary-historical construct „literary Bosnia“ which was based on the theory of the history of regional / provincial literature. By „literary Bosnia“ he meant everything that was its „provincial features“: folk history, genealogy, specific speech (dialect - ikavica), lifestyle (Muslims), and the canonical line consisted of Bosniak writers from Safvet-bega Bašagić, Musa Ćazim Ćatić, Edhem Mulabdić, Ahmed Muradbegović, to Alija Nametk, Enver Čolaković, Murat Šuvalić etc.Since in this period the pretensions towards Bosnia and Bosnian Muslims (Bosniaks) were also part of the Serbian national ideology, Jurčić's „literary Bosnia“ can be understood as a counterbalance to the then established Kršić's literary-historical construct „narrative Bosnia“. Unlike Kršić's „narrative Bosnia“, whose canonical line was mostly made up of Bosnian Serb writers (Ćorović, Kočić, Andrić, Ćopić, etc.), Jurčić's „literary Bosnia“ was made up of Bosniak writers as „the purest element of the Croatian people“.

Image segmentation has an important role in image processing and computer vision and it is widely used in numerous applications, including feature extraction, pattern recognition, scene analysis, object tracking. Due to its simplicity and effectiveness, multilevel thresholding approach to image segmentation has gained increased research attention in recent years. In this paper, the ability of two recently proposed metaheuristic algorithms, Honey badger algorithm and Chef-based optimization algorithm to ascertain the optimal threshold values based on Kapur’s entropy is systematically examined. The performance of the two multilevel thresholding image segmentation methods are assessed on a dataset of nine standard benchmark images. Based on a fixed number of independent runs, for each test image and a given number of thresholds, the multilevel thresholding performance is reported using the mean and standard deviation of Kapur’s entropy as well as the best objective function value and the associated threshold values.

Compared to conventional fire detection techniques, high-precision computer vision-based fire detection systems have a number of desirable characteristics, such as the ability to monitor large areas, provide a bountiful amount of information, and are easy to maintain. This paper extensively and systematically investigates the use of simple color-based rules for pixel-wise flame recognition in still images. The rules are evaluated on a hundred and nineteen test images that correspond to fires in urban environments. The performances of the considered flame recognition rules are reported in terms of Recall, Balanced accuracy, Accuracy, F1 score, and Matthews correlation coefficient. The best-performing rule is identified. More complex classifiers are formed by combining two or more simple rules. The experimental results show that simple color-based rules and some of their combinations can offer effective fire recognition performance.