• Open Access close
• CZ close
• Digital Humanities and Cultural Heritage close

Automatic translation of documents is an important task in many domains, in- cluding the biological and clinical do- mains. The second edition of the Biomed- ical Translation task in the Conference of Machine Translation focused on the au- tomatic translation of biomedical-related documents between English and various European languages. This year, we ad- dressed ten languages: Czech, German, English, French, Hungarian, Polish, Por- tuguese, Spanish, Romanian and Swedish. Test sets included both scientific publica- tions (from the Scielo and EDP Sciences databases) and health-related news (from the Cochrane and UK National Health Ser- vice web sites). Seven teams participated in the task, submitting a total of 82 runs. Herein we describe the test sets, participat- ing systems and results of both the auto- matic and manual evaluation of the trans- lations.

projektem OP VK PVBV – Popularizace vědy a badatelsky orientované výuky, reg. č. CZ.1.07/2.3.00/45.000

We summarize the involvement of our CEMI team in the ”NLI Shared Task 2017”, which deals with both textual and speech input data. We submitted the results achieved by using three different system architectures; each of them combines multiple supervised learning models trained on various feature sets. As expected, better results are achieved with the systems that use both the textual data and the spoken responses. Combining the input data of two different modalities led to a rather dramatic improvement in classification performance. Our best performing method is based on a set of feed-forward neural networks whose hidden-layer outputs are combined together using a softmax layer. We achieved a macro-averaged F1 score of 0.9257 on the evaluation (unseen) test set and our team placed first in the main task together with other three teams.

Neural natural language generation (NNLG) systems are known for their pathological outputs, i.e. generating text which is unrelated to the input specification. In this paper, we show the impact of semantic noise on state-of-the-art NNLG models which implement different semantic control mechanisms. We find that cleaned data can improve semantic correctness by up to 97%, while maintaining fluency. We also find that the most common error is omitting information, rather than hallucination. In Proceedings of INLG 2019, Tokyo, Japan

We present a test corpus of audio recordings and transcriptions of presentations of students' enterprises together with their slides and web-pages. The corpus is intended for evaluation of automatic speech recognition (ASR) systems, especially in conditions where the prior availability of in-domain vocabulary and named entities is benefitable. The corpus consists of 39 presentations in English, each up to 90 seconds long. The speakers are high school students from European countries with English as their second language. We benchmark three baseline ASR systems on the corpus and show their imperfection. Comment: SLSP 2019

In this paper, we introduce the system for inferring implicit computable knowledge from textual data by natural deduction. Our background system is Transparent Intensional Logic (TIL) with its procedural semantics that assigns abstract procedures known as TIL constructions to terms of natural language as their context-invariant meanings. The input data for our method are produced by the so-called Normal Translation Algorithm (NTA). The algorithm processes natural-language texts and produces TIL constructions. In this way we have obtained a large corpus of TIL meaning procedures. These procedures are furthermore processed by our algorithms for type checking and context recognition, so that the rules of natural deduction for inferring computable knowledge can be afterwards applied. Web of Science 24 1 48 29

Modern testing of thermophysiological comfort of functional textile fabrics involves not only their water vapour permeability and thermal resistance in dry state, but also testing of these properties in wet state. Besides stationary properties, also transient properties like warm-cool feeling or thermal absorptivity are important. In the paper, the concept of special version of this parameter, called moisture absorptivity, is explained and used for evaluation of transient moisture transfer (moisture management) between simulated wet human skin and selected dry functional underwear fabrics.

Magnetic Properties of a Novel CeCo0.715Si2.285 Compound J. Moud°ik, J. Prokle2ka, J. Pospi2il, V. Sechovský and I. Cisa°ova Charles University in Prague, Faculty of Mathematics and Physics, Department of Condensed Matter Physics, Ke Karlovu 5, 121 16 Prague 2, Czech Republic Charles University in Prague, Faculty of Science, Department of Inorganic Chemistry, Hlavova 2030/8, 128 43 Prague 2, Czech Republic

Traditional storytelling has been an essential constituent of Indigenous cultures in North America, including contemporary literary production. The perception of storytelling in Indigenous theories might, however, in some aspects differ from its Western literary critical understanding. This paper outlines the conceptual frameworks which shape and are reflected in Indigenous storytelling and examines relationality as the core principle of Indigenous knowledge system. Relationality does not only represent a frequent theme in Indigenous works, but it is also manifested in their narrative structure, and simultaneously can be used for their interpretation. In addition, the paper focuses on the equal role of the storyteller and the listener in Indigenous views on storytelling, the necessity to internalize a story to derive meaning, the power and animacy of stories and the importance of this for Indigenous typologies of stories, and on the issue of cultural appropriation, which stems from the different approaches to stories and knowledge as such in Indigenous and non-Indigenous perceptions.

Abstract A detailed morphological description of word forms in any language is a necessary condition for a successful automatic processing of linguistic data. The paper focuses on a new description of morphological categories, mainly on the subcategorization of parts of speech in Czech within the NovaMorf project. NovaMorf focuses on the description of morphological properties of Czech word forms in a more compact and consistent way and with a higher explicative power than approaches used so far. It also aims at the unification of diverse approaches to morphological annotation of Czech. NovaMorf approach will be reflected in a new morphological dictionary to be exploited for a new automatic morphological analysis (and disambiguation) of corpora of contemporary Czech.