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

On the 25th anniversary of the Dayton Agreement, this article assesses the current state of Euro-Atlantic integration in Bosnia and Herzegovina. Its starting point is that Dayton represented not a breach but a continuation of the politics which destroyed Yugoslavia and that salvaging the country’s democratic potential requires a paradigmatic break from virtually all its socio-political practices and institutions. In these contexts, the article’s central focus on the ‘Bosnian Spring’ which took place right across the country in (and in the run-up to) 2014 provides a salutary reminder of the ability of ordinary people to come together to demand socio-economic justice and, in doing so, to step out of the ethno-nationalist rigidities imposed by Dayton. While those protests ran out of steam, such events are the only ones by which elites can historically be persuaded to concede democratic ground. Dismantling the ethno-nationalist narrative remains the challenge and, when it becomes clear that change will not come from within the system, it becomes the task of ordinary citizens to create democratic institutions that are worth the label.

Continuous investment in education is a key factor in economic growth and development. Investing in education is an investment with a returns on investment and with multiple positive effects at the private and social levels. Earnings in the labor market represent private returns on investment in education and their measurement is in the center of the scientific literature. The main objective of this paper is to analyze modern theoretical and methodological approaches to measuring returns on investment in education, while analyzing the measurement methodology, indicators used and data sources. The scientific contribution of this paper is based on a concise and critical review of the scientific literature and trends in this field, with reference to the methodology used and indicators in measuring returns on investment in education.