Semantic Level Research Articles

Improving RGB-D SLAM accuracy in dynamic environments based on semantic and geometric constraints

Most current visual SLAM systems rely on static environment assumptions. However, in dynamic environments, the presence of dynamic objects can severely impair the performance of visual SLAM systems. To address this issue, we propose an algorithm that combines semantic segmentation and geometric constraints to eliminate the impact of dynamic objects on visual SLAM trajectory accuracy in dynamic indoor environments. First, a missed detection compensation algorithm based on the constant velocity model and region growing algorithm is proposed, which improves the accuracy of semantic or instance segmentation network and realizes accurate segmentation of dynamic objects at the semantic level in the new thread. Second, for the dynamic objects that cannot be recognized by the semantic segmentation network, a geometric constraint model based on the position correlation between feature points is presented to further eliminate the dynamic feature points that do not meet the motion consistency. Thereafter, the proposed method is integrated into the front-end of the ORB-SLAM2 system and evaluated using the RGB-D dataset. Experimental results demonstrate that the proposed method can improve the absolute trajectory accuracy of the ORB-SLAM2 system by an average of 96.90% in high dynamic scenes.

Integrating Sensor Ontologies with Niching Multi-Objective Particle Swarm Optimization Algorithm.

Sensor ontology provides a standardized semantic representation for information sharing between sensor devices. However, due to the varied descriptions of sensor devices at the semantic level by designers in different fields, data exchange between sensor devices is hindered. Sensor ontology matching achieves data integration and sharing between sensors by establishing semantic relationships between sensor devices. Therefore, a niching multi-objective particle swarm optimization algorithm (NMOPSO) is proposed to effectively solve the sensor ontology matching problem. As the sensor ontology meta-matching problem is essentially a multi-modal optimization problem (MMOP), a niching strategy is introduced into MOPSO to enable the algorithm to find more global optimal solutions that meet the needs of different decision makers. In addition, a diversity-enhancing strategy and an opposition-based learning (OBL) strategy are introduced into the evolution process of NMOPSO to improve the quality of sensor ontology matching and ensure the solutions converge to the real Pareto fronts (PFs). The experimental results demonstrate the effectiveness of NMOPSO in comparison to MOPSO-based matching techniques and participants of the Ontology Alignment Evaluation Initiative (OAEI).

Adapting Contingency Management for Hospitalized Patients with Stimulant Use Disorder.

Hospitals struggle to engage patients with stimulant use disorders, and little is known about how to adapt evidence-based behavioral interventions, such as contingency management (CM), for hospital settings. Our study is the first step in informing the design of a hospital CM intervention. We performed a qualitative study at a quaternary referral academic medical center in Portland, Oregon. We conducted semistructured qualitative interviews with CM experts, hospital staff, and hospitalized patients, eliciting input about hospital CM adaptations, anticipated challenges, and potential opportunities. We performed a reflexive thematic analysis at a semantic level and shared results for respondent validation. We interviewed 8 CM experts (researchers and clinicians), 5 hospital staff, and 8 patients. Participants felt CM could benefit hospitalized patients by supporting patient substance use disorder and physical health goals, especially by addressing the boredom, sadness, and loneliness of hospitalization. Participants emphasized that in-person interactions could improve patient-staff relationships by using "super positive" experiences to improve rapport. For successful hospital CM, participants emphasized CM core concepts and potential hospital adaptations, including identifying hospital-specific high-yield target behaviors, ensuring staff training, and using CM to support the hospital discharge transition. Participants also encouraged considering novel mobile app interventions, which may offer more flexibility in the hospital, recommending that such interventions include an in-person CM facilitator. Contingency management has potential to support hospitalized patients and improve patient and staff experience. Our findings can inform CM interventions for hospital systems seeking to expand access to CM and stimulant use disorder treatment.

Analysis of Arabic-Indonesian Translation Errors on Google Translate

This study aims to analyze the results of Google Translate machine translation in translating Arabic into Indonesian. The data of this study used primary data types in the form of literature from books with library sources were Arabic texts taken from the book Nurul Yaqin written by Sheikh Muhammad Al Khudari Bek and his translation books in Indonesian and taken by random sampling technique. This research uses a qualitative approach with a descriptive qualitative type. This study aims to analyze the results of machine translation google translate in translating Arabic into Indonesian. This research is related to translation, namely the process of translating from Arabic into Indonesian by looking at the process carried out by Google Translate and then analyzing the results of the translation. The primary data is in the form of Arabic texts taken from the book Nurul Yaqin written by Sheikh Muhammad Al Khudari Bek. Then the author takes researched from the subtitles in the book, namely, the cleavage of the breast of the Prophet Muhammad SAW. The focus of this research is to find out what translation errors are focused on the linguistic level, such as morphology, syntax, and semantics. In this study, the authors identified errors at the morphological, syntactical, and semantic levels contained in google translate from Arabic to Indonesian. By using error analysis (error analysis) with procedures, 1) collecting error samples, 2) identifying errors, 3) explaining errors, 4) classifying errors, and 5) evaluating errors. The results of the study show that in terms of the accuracy of vocabulary and terms, google translate has good translation quality. In terms of clarity and reasonableness, it means that Google Translate has not been able to convey ideas from the source language into the target language. Furthermore, grammatically, the results from Google Translate translation do not have grammar, then the results from Google Translate translation do not have a good grammatical structure and follow the rules that apply in the target Indonesian language. From these data, it can be seen that Google Translate should not be used as a basis for translating Arabic texts into Indonesian. A novice translator should prefer a dictionary instead of using Google Translate to try and improve their translation skills. So far, Arabic is still a difficulty for Indonesians, especially for students, and this difficulty is a linguistic problem that includes phonology, syntax, morphology, and semantics.

Exploring the nature of the gender-congruency effect: implicit gender activation and social bias.

The aim of the study was to explore the nature of the gender-congruency effect, characterized by a facilitation on the processing of congruent words in grammatical gender. Moreover, we explored whether resemblances between gender identities and gender attitudes with grammatical gender modulated lexical processing. We designed a gender-priming paradigm in Spanish, in which participants decided the gender of a masculine or feminine pronoun preceded by three different primes: biological gender nouns (mapping biological sex), stereotypical nouns (mapping biological and stereotypical information), and epicene nouns (arbitrary gender assignment). We found faster processing of gender congruent pronouns independently of the type of prime, showing that the grammatical gender feature is active even when processing bare nouns that are not conceptually related to gender. This indicates that the gender-congruency effect is driven by the activation of the gender information at the lexical level, which is transferred to the semantic level. Interestingly, the results showed an asymmetry for epicene primes: the gender-congruency effect was smaller for epicene primes when preceding the feminine pronoun, probably driven by the grammatical rule of the masculine being the generic gender. Furthermore, we found that masculine oriented attitudes can bias language processing diminishing the activation of feminine gender, which ultimately could overshadow the female figure.

An Efficient Data Analytics and Optimized Algorithm for Enhancing the Performance of Image Segmentation Using Deep Learning Model

Image segmentation is the key topic in computer vision and image processing with applications like robotic perception, scene understanding, video surveillance, image compression, medical image analysis, and augmented reality among many others. There are numerous algorithms are developed in the literature for image segmentation. This paper provides a broad spectrum of pioneering works for instance and semantic level segmentation with mask Region based Convolution Neural Network with Monarch butterfly Optimization (RCNN-MBO) architecture. The system is initially constructed in a Python environment with images of people and animals being input. Remove the unnecessary data from the gathered datasets during the pre-processing stage. Next, use a stochastic threshold function to segment the image. Then update the segmented images into a designed model for detecting and classifying a group of images. The main goal of the designed approach is to attain accurate prediction results also improve the performance of the designed model by attaining better results. To enhance the performance, two activation functions were used and MBO fitness is updated in the classification layer. It improves the prediction results and takes less time to detect and classify images. Finally, the experimental outcomes show the reliability of the designed approach by other conventional techniques in terms of accuracy, precision, sensitivity, specificity, F-measure, error rate, and computation time.

Semantic Challenges of Business Translation with Reference to English and Arabic: A Product-Oriented Approach

This study aims to investigate the various types of challenges and limitations at semantic levels that translators face in the business and economic translation products of tenders and bids from English into Arabic. To achieve the objectives of the study, a sample of four tenders of business documents was collected from a variety of national and international non-governmental organizations such as the bilingual specialized texts that were published between 2019 and 2020 in Iraq, specifically in Erbil, Kurdistan Region. The study relies on Holmes’ map (1988) of descriptive translation studies, developed by Toury (1995). To answer the research questions: how does a translator deal with specialised terminologies and expressions in translating business tenders and bids and what effect and role do semantic changes play in business and economic translation products? The study employs the analytical-descriptive translation methodology in analysing business texts. The analysis of business and economic texts has revealed that inaccurate synonyms at the semantic level, such as using different synonyms and translating semantic collocations, appear more frequently than other issues in translating business texts in the target language. The findings of the study have indicated that business and economic terms may have more than one accurate and faithful translation. Nevertheless, one translation can be appropriate and accurate. The study has concluded that cultural phenomena rarely appear in business and economic tenders and bids in comparison with other areas of economic and business texts, e.g., metaphors and idioms in general business texts.

Types of nostalgia and its causes

This study aims to investigate the various types of challenges and limitations at semantic levels that translators face in the business and economic translation products of tenders and bids from English into Arabic. To achieve the objectives of the study, a sample of four tenders of business documents was collected from a variety of national and international non-governmental organizations such as the bilingual specialized texts that were published between 2019 and 2020 in Iraq, specifically in Erbil, Kurdistan Region. The study relies on Holmes’ map (1988) of descriptive translation studies, developed by Toury (1995). To answer the research questions: how does a translator deal with specialised terminologies and expressions in translating business tenders and bids and what effect and role do semantic changes play in business and economic translation products? The study employs the analytical-descriptive translation methodology in analysing business texts. The analysis of business and economic texts has revealed that inaccurate synonyms at the semantic level, such as using different synonyms and translating semantic collocations, appear more frequently than other issues in translating business texts in the target language. The findings of the study have indicated that business and economic terms may have more than one accurate and faithful translation. Nevertheless, one translation can be appropriate and accurate. The study has concluded that cultural phenomena rarely appear in business and economic tenders and bids in comparison with other areas of economic and business texts, e.g., metaphors and idioms in general business texts.

Localization of Terms Associated with the Covid-19 Pandemic from English to Arabic: A Study of Linguistic Means and Machine Translation Techniques

The emergency conditions experienced by the world during the COVID-19 pandemic have led to radical changes in both the lexicographic and semantic levels of the Arabic language. As with other languages around the world, Arabic had to keep up with the rapid development and massive flow of information and to be able to localize the terminology associated with the pandemic in the English language as the dominant global language, especially in the scientific field. The period of the COVID-19 virus outbreak and its accompanying effects and developments witnessed the production of a huge number of newly coined terms in the English language, which posed a major challenge for Arabic translators and required them to search for linguistic means to localize these new terms, ranging from derivation and borrowing to translation and the use of machine translation techniques based on artificial intelligence programs."

Enabling End-Users in Designing and Executing of Complex, Collaborative Robotic Processes

Over the last years, capabilities of robotic systems have quantitatively and qualitatively improved. But going beyond isolated robotic systems, the integration and interoperability of robotic capabilities in complex work processes remains a major challenge. This lack of tools to integrate robots needs to be addressed on technical, semantic and organizational level. In the ROBxTASK research project, we developed an approach to support cooperation between different types of users in order to enable domain experts, with no robotic know-how, to work with robot-assisted workflows. By engineering robotic skills at a useful and usable level of abstraction for experts in different domains, we aim to increase re-usability of these skills on two different levels, (robotic) device level, and on level of application specific workflows. The researched prototype consists of a web platform, which allows (a) engineers to register (robotic) devices and the implemented skills of the devices, (b) domain experts to use a graphical task design environment to create workflows across multiple robotic devices and lastly (c) robot co-workers to download and execute the workflow code in a local environment with digital twins or real robots. Additionally skills and workflows can be shared across organisations. Initial user studies have shown that the visual programming environment is accessible and the defined skill-set is easy to understand even for domain experts that are inexperienced in the field of robotics.

Automatic Chinese knowledge-based question answering by the MGBA-LSTM-CNN model

The purpose of knowledge-based question answering (KBQA) is to accurately answer the questions raised by users through knowledge triples. Traditional Chinese KBQA methods rely heavily on artificial features, resulting in unsatisfactory QA results. To solve the above problems, this paper divides Chinese KBQA into two parts: entity extraction and attribute mapping. In the entity extraction stage, the improved Bi-LSTM-CNN-CRF model is used to identify the entity of questions and the Levenshtein distance method is used to resolve the entity link error. In the attribute mapping stage, according to the characteristics of questions and candidate attributes, the MGBA-LSTM-CNN model is proposed to encode questions and candidate attributes from the semantic level and word level, respectively, and splice them into new semantic vectors. Finally, the cosine distance is used to measure the similarity of the two vectors to find candidate attributes most similar to questions. The experimental results show that the system achieves good results in the Chinese question and answer data set.

Potentializing physical violenceHistorical Sociology of Polizei and Sozialstaat from a system theoretical perspective

Abstract Violence as a means of threatening the life and health of others is the foundation of power structures. According to a system theoretical understanding, however, physical violence becomes the basis for the power medium only when it is potentialized as a possibility for the political system. This potential threat contributes to the exodus of the political system through the formation of political organizations. This paper examines this hypothesis about violence and the political system from a historical, sociological perspective. I analyzed semantically the significance of Polizei, whose scope is said to have been limited from various governance practices, including welfare to the maintenance of security since the 19th century, in the development of labor protection policy, which was a precondition for the formation of the modern German welfare state. On the one hand, the work of the police in social policy was clearly distinguished from welfare on the level of political semantics; on the other hand, the appearance of other professionals made it possible for the police to continue to be involved in a wide range of social policy. These empirical findings show that the semantic differentiation of the political system with the potentialization of violence does not mean irrelevance and indifference to the environment; rather, it expands the scope of interest and makes it possible to relate to a more diverse range of subjects.

Oculomotor Referents of Reading Activity in Children with Dyslexia Aged 9–11

The problem of printed text processing impairments in dyslexics is of current interest but not well developed. The integration of the activity approach and eye-tracking methods provides us the opportunity to manage the system structural and functional analysis the reading as an complex activity. The purpose of the present study was to examine the oculomotor behavior as referent of mental actions performed in the process of reading a written text by children with dyslexia. The study of the oculomotor behavior during a text reading in 9–11year children with dyslexia was carried out. An eye movements during the text reading were recorded by means of eye-tracker. All participants were presented two kinds of texts: two expository texts and two narrative texts. The fixation’s number and duration and regressive saccade number and duration was registered; the distribution of regressive saccade with different amplitudes (very short, short, medium and long) were analyzed. Statistical analysis of data revealed the significant between group distinctions in all gaze movements measures. These data proved that children with dyslexia have reading activity disfunction both on the lexical and propositional and semantic levels.

A Multiview Text Imagination Network Based on Latent Alignment for Image-Text Matching

In image-text matching field, one of the keys to improving performance is to extract features with more semantic information. Existing works demonstrate that semantic enrichment through knowledge expansion can improve the performance. Most of them expand image features. However, the shortage of semantic information in text modality and the unilateral character of the view are often bottlenecks that limit the performance of image-text matching models. To solve the two problems, we aggregate knowledge from multiple views and propose a Word Imagination Graph (WIG). WIG can be used to expand textual semantic information by imagination based on input images. Then, utilizing WIG, we construct a novel Multi-View Text Imagination Network (MTIN). MTIN enables latent alignment of images and texts on tags which can assist matching on a semantic level. Results on Flickr30K and MS-COCO datasets can demonstrate the effectiveness of our method.

Валентность и сочетаемость в структуре англо- и русскоязычных многокомпонентных атрибутивных словосочетаний

The aim of the study is to determine the valence and collocational links and relations between the units of English-language and Russian-language multi-component attributive phrases. The scientific novelty of the study lies in systematising the theoretical and practical knowledge within the theory of the combination of linguistic units in relation to the analysis of multi-component attributive phrases in English and Russian, assessment of their valence potential and its manifestation through specific combinations. As a result, it has been found that valence and collocability in multi-component attributive phrases in the compared language formats are recorded primarily at the syntactic, lexical and semantic levels. Various forms of collocability emerge between the units of attributive construct with active and passive valence, such as formal and conceptual, arbitrary and conditioned, contact and distant, intrinsic and extrinsic, open and closed, absolute and relative, stylistic, emotionally expressive, associative proper, ontological and cognitive-semantic. By taking into account collocability, it is possible to penetrate deeper into the structure of multi-component attributive phrases and factor in their nature when analysing the linguocreative activity of the speaker.

MazeGen: A Low-Code Framework for Bootstrapping Robotic Navigation Scenarios for Smart Manufacturing Contexts

In this research, we describe the MazeGen framework (as a maze generator), which generates navigation scenarios using Grammatical Evolution for robots or drones to navigate. The maze generator uses evolutionary algorithms to create robotic navigation scenarios with different semantic levels along a scenario profile. Grammatical Evolution is a Machine Learning technique from the Evolutionary Computing branch that uses a BNF grammar to describe the language of the possible scenario universe and a numerical encoding of individual scenarios along that grammar. Through a mapping process, it converts new numerical individuals obtained by operations on the parents’ encodings to a new solution by means of grammar. In this context, the grammar describes the scenario elements and some composition rules. We also analyze associated concepts of complexity, understanding complexity as the cost of production of the scenario and skill levels needed to move around the maze. Preliminary results and statistics evidence a low correlation between complexity and the number of obstacles placed, as configurations with more difficult obstacle dispositions were found in the early stages of the evolution process and also when analyzing mazes taking into account their semantic meaning, earlier versions of the experiment not only resulted as too simplistic for the Smart Manufacturing domain, but also lacked correlation with possible real-world scenarios, as was evidenced in our experiments, where the most semantic meaning results had the lowest fitness score. They also show the emerging technology status of this approach, as we still need to find out how to reliably find solvable scenarios and characterize those belonging to the same class of equivalence. Despite being an emerging technology, MazeGen allows users to simplify the process of building configurations for smart manufacturing environments, by making it faster, more efficient, and reproducible, and it also puts the non-expert programmer in the center of the development process, as little boilerplate code is needed.

Iterative heterogeneous graph learning for knowledge graph-based recommendation

Incorporating knowledge graphs into recommendation systems has attracted wide attention in various fields recently. A Knowledge graph contains abundant information with multi-type relations among multi-type nodes. The heterogeneous structure reveals not only the connectivity but also the complementarity between the nodes within a KG, which helps to capture the signal of potential interest of the user. However, existing research works have limited abilities in dealing with the heterogeneous nature of knowledge graphs, resulting in suboptimal recommendation results. In this paper, we propose a new recommendation method based on iterative heterogeneous graph learning on knowledge graphs (HGKR). By treating a knowledge graph as a heterogeneous graph, HGKR achieves more fine-grained modeling of knowledge graphs for recommendation. Specifically, we incorporate the graph neural networks into the message passing and aggregating of entities within a knowledge graph both at the graph and the semantic level. Furthermore, we designed a knowledge–perceiving item filter based on an attention mechanism to capture the user’s potential interest in their historical preferences for the enhancement of recommendation. Extensive experiments conducted on two datasets in the context of two recommendations reveal the excellence of our proposed method, which outperforms other benchmark models.

Analisis Kesalahan Hasil Terjemahan Mesin Penerjemah Teks Bahasa Jerman ke dalam Bahasa Indonesia

Many online translation machines can translate more than 100 languages, including German into Indonesian and vice versa. However, the accuracy and standardization of the results still need to be investigated. This study aims to identify language errors in the translation of German texts into Indonesian using Google Translate and Bing Translator. This research is a descriptive qualitative research based on language error analysis. The data is derived from 12 German news articles published by Deutsche Welle. These German texts were translated into Indonesian using these two machine translators. The translated texts were analyzed for language errors at the semantic, morphological, and syntactic levels. The results of this study show that there are still language errors at the semantic, morphological, and syntactic levels in the translated texts produced by Google Translate and Bing Translator. Both translation machines tend to translate the text word for word. A translation result is an initial form of translation that needs to be edited to obtain an accurate translation.

Enhanced features in image manipulation detection

Image editing technology leads to tampering traces hidden in images hard to be identified, which results in great challenges for image manipulation detection. Therefore, it is necessary to enhance the characteristic response of tampering traces. In this paper, a two-stream network is exploited to extract tampering features from both semantic level and noise level through end-to-end training. At the input end of the network, a multi-scale high-frequency constrained convolution kernel module is proposed to extract multi-channel and multi-scale noise fingerprint information, which enhances the tampering traces. Moreover, we leverage multi-task learning to realize the classification of tampering techniques, the detection and segmentation of tampering regions. The tampered segmentation mask attention is introduced as a novel constraint to be fed back to the detection branch, which further enhances the tampering characteristics and improves the detection performance. Experiments on COVER, CASIA and NIST16 datasets prove that the proposed framework outperforms existing methods quantitatively and qualitatively. The proposed method effectively enhances the characteristics of tampered regions. Moreover, the proposed method has high robustness on JPEG recompression and image resize.

A study of the distribution of classes of lexical aspect in Spanish by Chinese learners of Spanish in oral narration

Abstract This paper aims to analyze the non-native use of Spanish verbs of different aspectual classes by Mandarin Chinese speakers performing oral narrative tasks. In this work, we adopt Vendler’s (Vendler, Zeno. 1967. Linguistics in philosophy. Ithaca: Cornell University Press) lexical-aspect classification (states, activities, accomplishments, and achievements), and we analyze the Spanish L2 acquisition of these categories by sinophones both at the level of the lexical form acquisition and at the semantic level. Based on data obtained from two oral narrative tasks elicited from storyboard vignettes performed by three groups of learners of L2/L3 Spanish (ranging from B1 to C1 levels of the CEFR), this work supports Jiang’s (Jiang, Nan. 2000. Lexical development and representation in a second language. Applied Linguistics 21(1). 47–77) L2 lexical acquisition pattern at the lexical form level, as our sinophone learners show a similar pattern to the one observed in native speakers in the use of aspectual class of compositional form (accomplishments)and aspectual classes of individual form (states, activities and achievements). However, at the lexical semantic level, we found that, besides L2 proficiency, the properties of lexical semantics may also impact learners’ L2 lexical processing.