• Digital Humanities and Cultural Heritage
• Conference object close
• Research close

Experience gained from commissioning and operation of three elliptical polarization undulators (EPU) at the TPS taught us that undulator driving systems can behave erratically following a beam dump or loss. In this work, we discuss possible harmful radiation sources in a storage ring and analyse the effect of lack of electronic component radiation resistance in the system. According to measurements of spatial radiation distribution at the TPS, we propose solutions and an improved design for Phase-II EPUs. Proceedings of the 10th Int. Particle Accelerator Conf., IPAC2019, Melbourne, Australia

Automatic prosodic event detection is important for both speech understanding and natural speech synthesis since prosody provides additional information over the short-term segmental features and lexical representation of an utterance. Similar to previous work, this paper focuses on automatic detection of coarse level representation of pitch accents, intonational phrase boundaries (IPB), and break indices. We exploit various classifiers and identify effective feature sets to improve performance of prosodic event detection according to acoustic, lexical, and syntactic evidence. our experiments on the Boston University Radio News Corpus show that the neural network classifier achieves the best performance for modeling acoustic evidence, and that support vector machines are more effective for the lexical and syntactic evidence. The combination of the acoustic and the syntactic models yields 89.8% accent detection accuracy, 93.3% IPB detection accuracy, and 91.1% break index detection accuracy. Compared with previous work, the IPB perfromance is similar, whereas the results for accent and break index detection are significantly better.

The paper deals with the transfer of finite verb structure from Tamil to Hindi using a rule-based approach in a machine translation system. This task is challenging as it involves the transfer of more grammatical information apart from lexical transfer and also because of the highly inflectional nature of finite verb morphology in Tamil compared to Hindi.

The rapid growth of online news belonging to different categories is causing users to spend a lot of time and effort searching for relevant and important news. Text categorization has a great significance in information retrieval and natural language processing where unstructured text can be organized into predefined categories. In this paper we investigate Uzbek news categorization using a convolution neural network and four word embedding models. We obtain two new word embeddings for Uzbek and present them in the Uzbek news categorization task.

Deep Learning techniques are powerful in mimicking humans in a particular set of problems. They have achieved a remarkable performance in complex learning tasks. Deep learning-inspired Neural Machine Translation (NMT) is a proficient technique that outperforms traditional machine translation tasks. Performing machine-aided translation on Indic languages has always been a challenging task considering their rich and diverse grammar. The neural machine translation has shown quality results compared to the traditional machine translation approaches. The task becomes problematic when it comes to low-resourced language like Sanskrit. This paper has presented our work on developing a dataset with Sanskrit as a language pair and an attention mechanism-based neural machine translation model trained on the Sanskrit-Hindi bilingual dataset. We have shown the significance of the attention mechanism to overcome the problem of long-term dependencies. The attention mechanism has shown promising results on low-resourced Indic languages.

Abstract Ensuring energy quality, i.e. compliance with specified or standardised energy performance indicators of buildings over the entire life cycle, can reduce the climate impact of the building sector. In this work, the causes for quality deficiencies in small building services installations as well as the structure of actors interacting with building services were analysed. On this basis, we developed a solution approach in the form of an information model including a catalogue of all requirements and the necessary content to address and improve the quality assurance of HVAC in small buildings. This information model will serve as the basis for future digital tools that can perform (partially) automated heating load calculations and the design of heating systems, improve installation logistics and installation quality through providing information and the necessary processing, and enable improved hydraulic balancing and control.