In the intense transport of bimodal contact load under high bed shear conditions, particles of the two fractions, which differ in size, tend to separate. The finer particles form an interfacial layer at the top of the mobile bed, while the coarser particles predominantly occupy the collisional layer above the interface. We report on tilting-flume experiments focusing on measurements of the local velocity and concentration distributions of particles for each fraction within the collisional layer. The results reveal that, due to vertical sorting, the presence of finer fraction particles rapidly diminishes with increasing elevation above the interfacial layer, whereas the local concentration of the coarser fraction is higher within the collisional layer and reaches a maximum near the center of the layer. The local velocity of the sediment is lower than the local velocity of water at the same elevations, indicating the expected local slip between the phases within the collisional layer. Furthermore, we utilize the measured integral quantities along with a previously collected dataset to compare similar flows with unimodal and bimodal transport. The comparison indicates that bimodal transport offers advantages in terms of reduced channel resistance and enhanced transport capacity.

The laboratory experiments on the intense transport of bimodal sediment were conducted in a tilted, glass-sided flume with a variable longitudinal slope. Two fractions of lightweight solids were used, primarily differing in particle size, and each had a distinct color. The observed solid-liquid flow exhibited characteristics of being steady, uniform, turbulent, and supercritical. The bimodal sediment was transported as a combined load, with the finer fraction primarily supported by carrier turbulence, and the coarser fraction supported by interparticle contacts in the transport layer above a plane surface of the bimodal stationary bed. Distributions of solids velocity and concentration were measured for each of the two fractions across the transport layer above the bed using optical methods employing high-speed cameras. Additionally, the distribution of carrier velocity was measured across the flow depth. The measurements revealed a non-uniform distribution of solids for both fractions, with the maximum concentrations at the top of the bed for the coarser fraction and within the transport layer for the finer fraction at the highest bed shear. The results of the measurements allowed for the identification of the degree of stratification in the high-concentration sediment-laden flow and facilitated the evaluation of the interaction between particles of different fractions in the transport layer at various elevations above the bed. Furthermore, they enabled the quantification of the proportion of particles of the two fractions in the total discharge of solids through the channel.

We present the results of laboratory experiments investigating the intense transport of bimodal bed load under high bed shear conditions in a tilting flume. Particles of two lightweight sediment fractions, differing in size, tend to separate during transport above the plane surface of an eroded mobile bed. Coarser fraction particles are predominantly present in the collisional layer above the bed, while finer fraction particles are primarily concentrated in the interfacial layer, which develops between the eroded bed and the collisional layer. This observed stratification of transported fractions influences their respective contributions to the total bed load discharge through the flume. Vertical distributions of local velocity and volumetric concentration were measured across the flow depth for each fraction separately, allowing the determination of each fraction's proportion in the total discharge. The experimental results were combined with a previously collected dataset to compare the discharges of bimodal and unimodal sediments under hydraulically similar conditions. Additionally, the experimentally determined discharges were evaluated against predictions from transport models designed for intense unimodal and bimodal bed loads.

An extended stereoscopic method, which identifies, and tracks particles based on their colour in solid-liquid flow, is tested for its capability to separately measure velocity distributions of particles of two fractions transported as bimodal sediment mixture in water flow through a laboratory flume. The principle of the tested method extension is a use of colour-based processing of images collected by two high-speed cameras which enables to filter out particles of one fraction from the image and leave particles of the other fraction in the image based on a selected colour hue range. The modified images are then processed by the original stereoscopic method to produce velocity distribution of particles of the individual fraction in the flow. The method extension is first tested in simple vertical flow carrying neutrally buoyant spherical particles of two distinct colours in a recirculation cell. In the next step, lightweight plastic particles of two fractions of different properties (size, shape, density) and colours are introduced to flow through a laboratory flume and velocity distributions of the two fractions are measured separately at flow conditions which mimic intense transport of bimodal combined-load in an open channel. Results exhibit a very good agreement with previous measurements with unimodal sediment in hydraulically similar flow.

Solid–liquid flows are encountered in various industrial and natural environments. The internal structure of such flows is highly sensitive to the grading of the solid particles present. In this experimental study, an extended stereometric method is employed to assess the distributions of velocity of particles of different fractions, distinguished by different colors, in vertical and nearly horizontal granular flows. In the vertical flow experiments, mixtures comprising three fractions of lightweight particles, characterized by a very similar density, size, and shape, were tested. The results affirmed the method’s ability to discern particle velocity differences on the order of millimeters per second, establishing its suitability for characterizing nearly horizontal open-channel flows with bimodal mixtures that are stratified and exhibit more complex velocity distributions. Tilting flume experiments, incorporating additional measurements of water velocity distribution, allowed for the evaluation of local slip between water and particles, as well as between particles of the two fractions in the flow. The results indicated that, although the local slip velocity was relatively small, the average slip velocity between the carrying water and transported particles was significantly larger. This factor must be taken into consideration when evaluating bed friction or bed erosion for granular flow in a channel with an erodible bed.

In this contribution, the effects of solids segregation on intense transport of solids and bed friction in a laboratory flume are evaluated by comparing bimodal solids flows with the corresponding flows of unimodal solids. The comparison is carried out for two types of solids transport: contact-load transport and combined-load transport. Experiments with intense transport of bimodal solids mixtures composed of two solids fractions that differ primarily in particle size showed a process of vertical sorting, resulting in partial segregation of the fractions in a flume. The segregation affected a layered structure of flow above the plane surface of an eroded bed. Lightweight solids fractions of different colors were used in the experiments to enable clear visual observation of the segregation. Measurements of integral quantities of steady, uniform solid-liquid flow were used to quantify solids transport parameters and bed friction parameters. The observed segregation patterns differ for the two types of solids transport and so differ their effects on solids discharge and bed friction coefficient in flows of the two bimodal mixtures.

Steady uniform open-channel flow with intense transport of combined load (suspended load and contact load) above plane mobile bed is modelled using a new developed transport model based on the kinetic theory of granular flow. The model is an extension of the recently developed model for collisional transport and the new model employs the mixing-length concept to incorporate a contribution of flow turbulence in supporting transported sediment particles additional to the dominating support contribution by interparticle contacts. The model predicts flow rates of water and sediment for a given combination of the channel depth and longitudinal slope and includes predictions of the solids distribution and flow velocity across the flow depth. Validation experiments with lightweight sediment in a laboratory tilting flume provide results for a comparison with new model predictions. The experiments include measurements of the distribution of the local sediment concentration using the camera-based laser stripe technique and measurements of the distribution of velocity of solids using the imaging technique based on Particle Tracking Velocimetry with two synchronized high-speed cameras. Furthermore, local velocities of water are obtained from Pitot-tube measurements. The paper presents results of the transport model and their comparison with results of the laboratory experiment for a relatively broad range of bed shear conditions (Shields parameter between 0.8 and 1.8 approximately). A discussion includes criteria and conditions for the local support of particles by turbulent eddies in the transport layer of the flow transporting combined load.

This paper discusses two-phase laminar flows of mixtures of solids in non-Newtonian carrier consisting of spherical particles in a 50-mm pipe at the Czech Technical University in Prague. Special attention is paid to the frictional head loss and thickness of transport layers. For the prediction of various transport characteristics, a layered model is used. Suitability of the layered model to predict the transport characteristics is investigated. Prediction results are compared to own experimental data.

The contribution deals with the determination of the ratio of volumes of particles of individual fractions in flows transporting bimodal granular mixtures. The spatial distribution of the ratio is obtained by analyzing images collected by video recording. The measuring method and post-processing analysis are demonstrated on experiments in our new recirculating fluidization cell. The tested bimodal mixture is composed of two granular fractions of different colors – black and white. The analysis is based on a local identification of particles and their background in the collected images. More precisely, the identification of individual particles is based on the grayscale screening of the images. The method allows to identify both the instantaneous and time-averaged spacious distribution of the grain volume ratio. The experimental results reveal that the measuring method is successful if constant light condition is secured and the medium carrying particles is sufficiently transparent. The method is applicable to laboratory flows with sediment transport no matter whether sediment particles are transported as bed load or as suspended load.

Experimental investigation of the effect of particle geometry on the values of hydraulic conductivity using 5 plastic particle fractions of different shapes and sizes is presented in this paper. Conducted experiments serve as a background for ongoing research focused on sediment transport of erodible bed in tilting flume. When a flowing mixture of water and particles evolves above a stationary bed, the determination of flow rate of water filtered through stationary bed is very important in order to estimate the inaccuracy of total flow rate trough the flume. In order to evaluate the amount of infiltrate water, experiments were carried out in both laminar and turbulent flow regimes. The transition area was identified, and the results show only small effect of particle geometry on the transition between regimes. Experimentally identified values of hydraulic conductivity were used to calculate the amount of infiltrate water in the tilting flume for different inclination angles. Experimentally identified values of hydraulic conductivity are, for example, useful in designing industrial water filters, as the materials used to conduct experiments and those in filters are similar.

We generalize the word analogy task across languages, to provide a new intrinsic evaluation method for cross-lingual semantic spaces. We experiment with six languages within different language families, including English, German, Spanish, Italian, Czech, and Croatian. State-of-the-art monolingual semantic spaces are transformed into a shared space using dictionaries of word translations. We compare several linear transformations and rank them for experiments with monolingual (no transformation), bilingual (one semantic space is transformed to another), and multilingual (all semantic spaces are transformed onto English space) versions of semantic spaces. We show that tested linear transformations preserve relationships between words (word analogies) and lead to impressive results. We achieve average accuracy of 51.1%, 43.1%, and 38.2% for monolingual, bilingual, and multilingual semantic spaces, respectively.

In this paper we evaluate our new approach based on the Continuous Bag-of-Words and Skip-gram models enriched with global context information on highly inflected Czech language and compare it with English results. As a source of information we use Wikipedia, where articles are organized in a hierarchy of categories. These categories provide useful topical information about each article. Both models are evaluated on standard word similarity and word analogy datasets. Proposed models outperform other word representation methods when similar size of training data is used. Model provide similar performance especially with methods trained on much larger datasets.

In this paper we extend Skip-Gram and Continuous Bag-of-Words Distributional word representations models via global context information. We use a corpus extracted from Wikipedia, where articles are organized in a hierarchy of categories. These categories provide useful topical information about each article. We present the four new approaches, how to enrich word meaning representation with such information. We experiment with the English Wikipedia and evaluate our models on standard word similarity and word analogy datasets. Proposed models significantly outperform other word representation methods when similar size training data of similar size is used and provide similar performance compared with methods trained on much larger datasets. Our new approach shows, that increasing the amount of unlabelled data does not necessarily increase the performance of word embeddings as much as introducing the global or sub-word information, especially when training time is taken into the consideration.

Croatian is poorly resourced and highly inflected language from Slavic language family. Nowadays, research is focusing mostly on English. We created a new word analogy corpus based on the original English Word2vec word analogy corpus and added some of the specific linguistic aspects from Croatian language. Next, we created Croatian WordSim353 and RG65 corpora for a basic evaluation of word similarities. We compared created corpora on two popular word representation models, based on Word2Vec tool and fastText tool. Models has been trained on 1.37B tokens training data corpus and tested on a new robust Croatian word analogy corpus. Results show that models are able to create meaningful word representation. This research has shown that free word order and the higher morphological complexity of Croatian language influences the quality of resulting word embeddings.

Semantic textual similarity is the core shared task at the International Workshop on Semantic Evaluation (SemEval). It focuses on sentence meaning comparison. So far, most of the research has been devoted to English. In this paper we present first Czech dataset for semantic textual similarity. The dataset contains 1425 manually annotated pairs. Czech is highly inflected language and is considered challenging for many natural language processing tasks. The dataset is publicly available for the research community. In 2016 we participated at SemEval competition and our UWB system were ranked as second among 113 submitted systems in monolingual subtask and first among 26 systems in cross-lingual subtask. We adapt the UWB system for Czech (originally for English) and experiment with new Czech dataset. Our system achieves very promising results and can serve as a strong baseline for future research.

The word embedding methods have been proven to be very useful in many tasks of NLP (Natural Language Processing). Much has been investigated about word embeddings of English words and phrases, but only little attention has been dedicated to other languages. Our goal in this paper is to explore the behavior of state-of-the-art word embedding methods on Czech, the language that is characterized by very rich morphology. We introduce new corpus for word analogy task that inspects syntactic, morphosyntactic and semantic properties of Czech words and phrases. We experiment with Word2Vec and GloVe algorithms and discuss the results on this corpus. The corpus is available for the research community.

Restaurant Reviews CZ ABSA - 2.15k reviews with their related target and category The work done is described in the paper: https://doi.org/10.13053/CyS-20-3-2469

This paper describes our system used in the Aspect Based Sentiment Analysis (ABSA) task of SemEval 2016.Our system uses Maximum Entropy classifier for the aspect category detection and for the sentiment polarity task.Conditional Random Fields (CRF) are used for opinion target extraction.We achieve state-of-the-art results in 9 experiments among the constrained systems and in 2 experiments among the unconstrained systems.

We present our UWB system for Semantic Textual Similarity (STS) task at SemEval 2016. Given two sentences, the system estimates the degree of their semantic similarity. We use state-of-the-art algorithms for the meaning representation and combine them with the best performing approaches to STS from previous years. These methods benefit from various sources of information, such as lexical, syntactic, and semantic. In the monolingual task, our system achieve mean Pearson correlation 75.7% compared with human annotators. In the cross-lingual task, our system has correlation 86.3% and is ranked first among 26 systems.