Sentiment analysis on social media is increasingly challenging due to the growing prevalence of multimedia posts to convey complex emotions. Traditional sentiment analysis methods, which rely solely on textual or visual information, often fail to capture the nuanced interplay between modalities, leading to suboptimal predictions. Motivated by the limitations of unimodal approaches in handling sarcasm, this study proposes a novel multimodal sentiment analysis model that integrates textual and visual features for enhanced understanding. The proposed model leverages BERT and BiLSTM networks for capturing contextual and sequential dependencies in text, while CNNs extract rich spatial features from images. An attention-based fusion mechanism further enhances the integration of these modalities by focusing on the most informative elements in both text and pictures. Pretrained CNNs are fine-tuned through transfer learning to improve visual feature extraction, and the combined multimodal representation is used for sentiment prediction. Experimental results demonstrate that the proposed model achieves 92% accuracy, 89.5% precision, 89% recall, and an F1-score of 89.2%, outperforming conventional unimodal and existing multimodal approaches. This work lays the foundation for future extensions to include other modalities such as audio or video, enabling more comprehensive emotion understanding in real-world applications.

This article examines the general paradigm of language tools used in benevolence, with an emphasis on their functional and semantic features. Benevolence is considered as a special type of speech acts that perform an important communicative, cultural and ethical function in society. The author analyzes phrases and phraselogisms in which various aspects of the concept of “benevolence” are manifested – benevolence, respect, wishes for success, health, longevity, etc. It is noted that these linguistic units not only convey the emotional and evaluative attitude of the speaker, but also reflect the mentality, spiritual values and ethnocultural specifics of the people. Special attention is paid to the analysis of semantic fields and grammatical structures typical for expressing well-wishes in oral and written speech. It is revealed that phraseological units in this area represent stable cultural models that form nationally determined forms of speech etiquette and interpersonal communication. Thus, good wishes act as a significant element of the linguistic picture of the world, connecting the language, culture and worldview of the people.

ABSTRACT In the space between long-distance photometric observations and close-range visual feature observations, images that present only silhouette information of small bodies are valuable for estimating the physical characteristics and establishing a body-fixed coordinate frame. The voxel-based shape-from-silhouette (SFS) method has been proven to be able to estimate the shape and rotation axis of small bodies. However, the large number of voxels and iterations causes the SFS method to be inefficient and time-consuming, and it cannot meet the requirements for the autonomy and timeliness of on-orbit missions. In addition, existing studies on the silhouette-based rotation axis estimation lack the exploration of adaptability to multiple small body observation conditions. In this study, we developed a fast and effective integrated inversion method for both shape and rotation axis, which is suitable for images where small bodies occupy only tens to hundreds of pixels. First, we developed an octree-based voxel-divided SFS (VD-SFS) algorithm to enable the establishment of the silhouette voxel model hierarchically and efficiently. Then, by defining the rotation model of the small body and applying the VD-SFS algorithm, the rotation axis orientation can be estimated according to the silhouette similarity between the silhouette model and the actual images. Finally, simulation data under different observation conditions verify that the proposed method can significantly enhance the modeling efficiency, reducing both modeling time and memory consumption by several times compared to traditional methods. Additionally, the method proves effective in estimating the rotation axis of the small body, and remains robust even when images have a high sun phase angle. Using data from the approach phases of the Rosetta and OSIRIS-REx missions, the rotation axis error was estimated at under 2° for 67P and around 5° for Bennu, demonstrating that the proposed method can be effectively applied to on-orbit missions.

BackgroundAccurate automatic speech recognition (ASR) is essential for the objective analysis of speech during cognitive testing. Here, we describe the performance of seven individual ASR engines and selected consensus ASR (CASR) engine combinations when analyzing tokens generated by 453 randomly selected older subjects (mean age 60.5 years, 52% non‐white) who completed normative enrollment testing for the California Cognitive Assessment Battery (CCAB).MethodHigh‐quality (24‐bit 48 kHz) digital recordings were captured via noise‐canceling head‐mounted microphones during at‐home testing. Recordings were downsampled and transcribed using seven ASR engines: Amazon Transcribe, Google Cloud, Microsoft Azure, IBM Watson, Rev.ai, Vosk, and UWP (Windows real‐time transcription). This analysis focuses on three discursive speech tasks: two logical memory stories and a picture description with delayed recall. Custom grammars consisting of the words in the logical memory stories and key picture elements were used with each engine. Transcripts containing 412,741 words were reviewed and corrected using the CCAB's transcript review tool (Figure 1). The evidence for each utterance was estimated from lookup tables relating transformed engine confidence to accuracy for each ASR engine. All speech—including extraneous utterances (e.g., “I forgot that!!”) and subvocalizations—was retained. Word‐level evidence was estimated from lookup tables mapping transformed engine confidence values to empirical accuracy for each ASR engine. For CASR, word alignments were computed using the Levenshtein algorithm, and the total evidence (including for null entries and dysfluencies) was summed across engines to identify the most probable consensus word.ResultWord error rates (WERs) for single ASR engines ranged from 3.68% to 47.67%. CASR analysis incorporating all engines reduced the WER to 2.12% (Table 1). In addition, CASR‐derived confidence scores correlated more strongly with transcription accuracy (r = 0.75) than those of the best individual engine (r = 0.47) and more accurately predicted transcription errors (Figure 2). CASR also mitigated transcription biases associated with race and gender.ConclusionCASR achieves near‐human transcription accuracy, flags potentially unreliable transcripts, and reduces demographic bias, demonstrating its utility and robustness in high‐stakes cognitive assessments.

Aim. The identifying how human attitudes towards animals, which are an important element of the German value picture of the world, are reflected in the language at the lexical level and form the axiological potential of words. Methodology. Based on lexicographic sources and texts from the German-language media space, a corpus of empirical material was compiled, including lexical units with the component “SCHWEIN”, and analyzed using the methods of component analysis, discourse analysis and the modeling method. Results. It has been proven that the anthropocentrism of the language system is manifested in the semantics of words and phraseological units with the component “SCHWEIN”; a model for describing the meaning of zoolexemes has been developed, comprising three components: animalistic, anthroponymic, and a relational component; a systematization of German lexical units with the “SCHWEIN” component based on the content of the identified relationships is proposed, including five groups; it has been established that the direction of the relations – emphasizing the benefit for humans or animals – determines the pragmatic potential of animalistic lexical units. Research implications of this study lies in the development of animalistic themes within the framework of a new interdisciplinary direction – ecolinguistics. It also contributes to the expansion of existing methods of describing zoolexemes and the application of the developed model to analyze the semantic and pragmatic features of other zoolexemes. In addition, the study enriches our understanding of the German ethnic group's value system.

This study investigates how the method used by participants to assess the beauty of pictures influences their preference for the compositional rules of symmetry, balance, and proximity. The hypothesis that production methods (actively arranging picture elements) prompt a local perspective, favoring proximity, while evaluation tasks (rating precomposed pictures) elicit a global perspective, favoring symmetry and balance, was tested in two experiments. Experiment 1 demonstrated that (positional) symmetry was preferred over balance, and balance over proximity, when participants rated precomposed pictures. Experiment 2, employing a production method with movable elements, showed a frequent use of proximity, yet also a tendency toward (positional) symmetry. The combined results indicate that assessment methods substantially impact the preferred composition rules.

This article is devoted to the reconstruction of a fragment of the cognitive level of the linguistic personality of R. I. Fraerman, a Soviet writer, namely, the reconstruction of the concept of the Far East based on the material of the work "Wild Dog Dingo, or The Story of First Love". The subject of the study is the linguistic representation of a concept that is significant for the entire linguistic picture of the world by R. I. Fraerman. In modern linguistics, a number of definitions of the concept and approaches to its study, description and analysis coexist. In this article, the concept is understood as a mental representation of a person's knowledge of reality, carrying cultural information and finding its expression in various sign systems. This definition of the concept belongs to researchers in the field of conceptology O. A. Alexandrov and O. A. Andreeva. The study is based on the methods and techniques of semantic and cognitive description of the concepts of Z. D. Popova and I. A. Sternin. In addition to the methods and techniques of semantic and cognitive description of concepts, the study uses a descriptive method, including methods of observation, comparison, classification, the method of component analysis and individual techniques of statistical analysis. The scientific novelty of the research is determined primarily by the choice as its subject of the description of the concept that makes up the individual linguistic picture of the world of R. I. Fraerman, whose work, including the linguistic aspect, is currently completely unexplored. The following conclusions were obtained during the study: 1) the concept of the Far East as an integral element of the entire linguistic picture of the writer's world is evaluative; 2) most of the nominations representing the concept name certain natural phenomena of the region (its flora, fauna, relief, climate) and features, phenomena of the life of the indigenous people (Nanai); 3) vocabulary included in the nominative the field is quite detailed and specific, which is due to the writer's personal acquaintance with the specified region (the Far East).

Although dating apps are the preferred means of meeting sexual and romantic partners, users frequently experience disappointment, highlighting the importance of understanding self-presentation strategies and selection processes to mitigate negative experiences. This study examines self-presentation strategies in Tinder profile pictures, aiming to identify typological patterns, characterize common profile-building strategies, and analyze differences by age, gender, and sexual orientation. We employed a mixed dimensional approach-both categorical and numerical-to characterize and categorize 1000 Tinder profile pictures. A descriptive category set was developed to analyze key picture elements, and a not-safe-for-work nudity index was computed using an open-source neural network. We then applied K-means clustering to identify patterns in the data. The main outcome measures included the clustering distribution of profile picture types and their associations with demographic variables (Standardized Pearson Residuals). Our analysis identified nine prototypical Tinder profile image categories, confirming the presence of stereotypical patterns in self-presentation. Additionally, we found statistically significant associations between profile clustering and user demographics, particularly age, gender, and sexual orientation. Stereotyped self-disclosure in dating apps may hinder partner selection by reinforcing social biases related to age, gender, and sexual orientation, with potential consequences for sexual and couple therapy. Key strengths include the use of a large and diverse dataset, robust cluster validation techniques, and a novel approach to analyzing self-presentation on dating apps. Limitations, however, include potential biases due to the categorical nature of the data, difficulties in capturing individual nuances in self-presentation, and the inability to account for Tinder algorithm influences on image use. Clustering techniques provide an empirical framework for identifying stereotypical self-presentation patterns and demographic differences, which could be extended to analyzing written descriptions and partner selection mechanisms.

ABSTRACT Digital Elevation Models (DEMs) are extensively utilized for terrain analysis, representation, and visualization. Various application scenarios require DEMs at different Level of Details (LODs). Although traditional multi-scale landform expression methods primarily target DEMs at small scales, the simplifying Digital Surface Models (DSMs) for urban modeling at large scales remains a significant challenge. By integrating map generalization theory with computer vision techniques, we developed a novel method for urban building simplification in DSMs, termed Building Simplification in Digital Surface Models (BS-DSM). First, the buildings extracted from the DSMs are subjected to morphological analysis and aligned to a specific orientation. Next, the DSMs are divided into rectangular pixel blocks through energy-driven sampling, followed by horizontal simplification of building shapes in a 2D projection plane according to the geometric characteristics of these pixel blocks. Finally, to preserve the average height and total volume of the simplified 3D buildings in the vertical direction, the building heights across different pixel blocks are adjusted and interpolated based on the skeleton lines of building roofs and calculations of adjacent height values. The proposed BS-DSM method was evaluated using the publicly available Vaihingen DSM dataset. The result shows that the BS-DSM method performs better in simplifying building shape and height while meeting basic multi-scale expression constraints compared with traditional filtering methods.

I propose an approach to liar and Curry paradoxes inspired by the work of Roger Swyneshed in his treatise on insolubles (1330-1335). The keystone of the account is the idea that liar sentences and their ilk are false (and only false) and that the so-called “capture” direction of the T-schema should be restricted. The proposed account retains what I take to be the attractive features of Swyneshed’s approach without leading to some worrying consequences Swyneshed accepts. The approach and the resulting logic (called “Swynish Logic”) are non-classical, but are consistent and compatible with many elements of the classical picture including modus ponens, modus tollens, and double-negation elimination and introduction. It is also compatible with bivalence and contravalence. My approach to these paradoxes is also immune to an important kind of revenge challenge that plagues some of its rivals.

This work illustrates a novel application of a supervised superpixel-based segmentation method for root micrograph classification and total fungal colonization rate estimation. Two procedures relying on successive classifier application on different root micrographs or on the same micrograph but with an increasing number of labels to be assigned to each picture element category are compared to a reference grid-intersect count method. Finally, supervised classification with at least 16 labels on the same picture appears as a convenient method for obtaining rapid and confident colonization rate estimates. We suggest this kind of method may be easily and routinely implemented for research or educational purposes.

This research focuses on the ideological sphere of criminals with the highest status in the Russian Federation. This ideological sphere was studied in literary sources of various kinds on the basis of repeatability (the existence of linguistic facts) and averaging (external and internal confrontation of sources). It is suggested that, in speech, there exist some selective overinterpretations of world religions that neglect basic elements of the traditional law-abiding picture of the world and that are directly based on literary fiction instead of the scientific literature. On the other hand, there can be some search for faith connected with the belief in spiritual knowledge from the dead, divine beings, and God.

We sought to evaluate the ability of automated speech and language features to longitudinally track fluctuations in the major psychosis domains: Thought Disorder, Negative Symptoms, and Positive Symptoms. Sixty-six participants with psychotic disorders were assessed soon after inpatient admission, at discharge, and at 3- and 6-months. Psychosis symptoms were measured with semi-structured interviews and standardized scales. Recordings were collected from paragraph reading, fluency, picture description, and open-ended tasks. Relationships between psychosis symptoms and 357 automated speech and language features were analyzed using a single component score and as individual features, using linear mixed models. We found that all three domains demonstrated significant longitudinal relationships with the single component score. Thought Disorder was particularly related to features describing more subordinated constructions, less efficient identification of picture elements, and decreased semantic distance between sentences. Negative Symptoms was related to features describing decreased speech complexity. Positive Symptoms domain score did not show relationships with individual features that survived p-value correction, but Suspiciousness was related to decreased use of nouns and Hallucinations was related to greater semantic distances. These relationships were largely robust to interactions with gender and race. Interactions with timepoint revealed variable relationships during different phases of illness (acute vs. stable). In summary, automated speech and language features show promise as scalable, objective markers of psychosis severity. Detailed attention to clinical setting and patient population is needed to optimize clinical translation.

The article is devoted to the study of one of the key elements of the structure of the criminalistic characteristic – the trace picture of a criminal offense, which is formed as a result of leaving a young person in danger. The author empirically confirmed that the trace picture of leaving a minor in danger provides criminalistically significant information that is necessary for a successful pre-trial investigation. In particular, it contributes to the formulation of substantiated versions of the wrongful act, the appointment of forensic examinations and the conduct of a set of investigative (search) actions necessary to qualify the crime and bring the guilty person to justice. According to the results of the analysis of scientific sources devoted to the study of the trace picture as an element of the criminalistic characteristic associated with the act of leaving a minor in danger, the approaches that most fully characterize criminal offenses of the category under study are determined. The author notes that the trace picture of leaving a young person in danger is a complex of material, ideal and digital (information) traces. The article offers definition and discloses the constituent elements of the trace picture, which are most inherent in the crimes of the studied category. On the basis of the analysis of the materials of criminal proceedings of the studied category, the constituent elements of the trace picture of the commission of leaving a minor in danger are described. According to the results of the study, correlation relationships of the trace picture of the crime with other elements of the forensic characteristic, in particular by the method, the situation of the criminal offense and the identity of the criminal, have been substantiated. In the article, the author proves that the method of committing a crime is the main factor that determines the trace picture. The results of the study have practical value for law enforcement agencies in the context of preventing criminal offenses related to leaving minors in danger. In addition, they can be used in scientific developments of investigation methods for certain types of criminal offenses and further scientific research.

Abstract The question of what meaning music conveys is an ancient and unresolved issue. Central to discussions have been the relation between meaning in language and meaning in music: In spite of obvious differences, the challenge remains of providing an account that could place the two areas as elements of the same overall picture. This paper suggests an approach based on advances in functional-cognitive linguistics on the one hand and cognitive semiotics on the other. From functional-cognitive linguistics it draws on two key elements: (1) a division between three sites of meaning, rather than a monolithic approach, recognizing that meaning in actual usage events, in individual brains, and in society cannot be reduced to one thing. (2) A division in actual usage events between ‘input’ and ‘meaning construction’, stressing the constitutive role of the recipient in understanding the nature of meaning. From cognitive semiotics it draws on the concept of the semiotic hierarchy. The key point is the understanding of the nature of meaning as a broader and more fundamental property than found in the linguistic tradition.

Abstract Searching for primordial gravitational waves in the cosmic microwave background (CMB) polarization signal is one of the key topics in modern cosmology. Cutting-edge CMB telescopes require thousands of pixels to maximize mapping speed. Using a modular design, the telescope focal plane is simplified to several detector modules. Each module has hundreds of pixels including antenna arrays, detector arrays, and readout arrays. The antenna arrays, as the beam defining component, determine the overall optical response of the detector module. In this article, we present the developments of 6 inch broadband antenna arrays from 80 to 170 GHz for the future IHEP focal plane module. The arrays are fabricated from 42 6 inch silicon wafers including 456 antennas, 7% more pixels than the usual design. The overall in-band cross polarization is smaller than −20 dB and the in-band beam asymmetry is smaller than 10%, fulfilling the requirements for primordial gravitational wave search.

The classification and evaluation of orbital weld bead data using Machine Learning techniques provides a novel way to enhancing weld quality assessment processes. In this study, the random forest machine learning method was used to categorize weld quality using digital pictures of weld beads from 50 samples. The random forest model was trained utilizing essential picture elements to distinguish between acceptable and unsatisfactory weld quality, allowing for accurate categorization. The model's performance was evaluated using three categorization indicators: classification accuracy, F1-score, and the area under the receiver operating characteristic (ROC) curve (AUC). The RF model has a classification accuracy of 93.5%, an F1-score of 92.8%, and an AUC of 0.96, indicating its great dependability and robustness in determining weld bead quality. Feature selection and hyperparameter optimization improved the model's capacity to distinguish between high-quality and faulty welds. This work highlights the possibility for incorporating machine learning techniques, notably the random forest algorithm, into automated weld quality evaluation systems. The random forest model's great performance in categorizing weld quality demonstrates its usefulness in industrial settings, providing a path to increased efficiency and precision in defect identification and quality assurance procedures. Key Words: random forest, hyperparameters, machine learning, categorization, feature selection

We previously introduced the concepts of optical centroid efficiency (OCE) and enclosed energy within a rectangular pixel (EOD). We applied them to an ideal lens with and without a central obscuration for two different detector pixel sizes. Also, we analyzed the performance of OCE vs. EOD for the following three Seidel primary aberrations of an ideal lens: spherical, coma, and astigmatism, plus defocus. In this paper, we concentrate on three different optical remote sensing instrument configurations. We burden them with a set of aberrations to mimic realistic generalized error budgets that cover potential ground, lunch, and on-orbit environmental conditions. The shape of the OCE vs. EOD curve depends to a large degree on the dominant aberration. With the proper choice of detector pixel size, OCE increases with EOD when EOD is larger than 0.6. The increased detector pixel size is advantageous for structures that enhance diffraction effects, and for off-axis and asymmetrical configurations. Analytical and experimental tests are proposed for original critical cases. Furthermore, OCE and EOD, as functional figures of merit, may be effectively applied to instruments for monitoring tumors and their evolution to cancerous tissue, leading to timely diagnosis.

The discovery of veins in the forearm continues to be a significant obstacle, especially in susceptible patient groups such as the elderly, youngsters, and those who are obese. Conventional approaches typically fail to find veins effectively on the very first try. Not only does this make the patient uncomfortable, but it also reduces the effectiveness of the therapeutic procedure. To solve this ongoing problem, we suggest a cutting-edge method that uses a U-Net model coupled with ResNet (Residual Network). This technique is intended to considerably improve the accuracy of vein recognition. Because of the addition of ResNet, the model's capability to learn detailed elements in medical pictures is much enhanced, which results in a significant improvement in its performance. The performance of the proposed method achieved 96% accuracy when applied to a dataset of forearm medical photos, an 80% precision rate, and a minimum loss function of 0.06. These results were achieved through the application of the Enhanced Forearm Vein Detection method. By demonstrating that the model can outperform conventional approaches, these findings indicate that the model represents a significant advancement in the field of medical picture segmentation and vein recognition. The use of this cutting-edge method provides medical practitioners with a revolutionary instrument that guarantees a venous access procedure that is quick, accurate, and less invasive

The article presents a linguocultural analysis of the connotative meanings of ergonyms in St. Petersburg, particularly the names of restaurants and cafes. The subject of the study is the connotative meanings of ergonyms in St. Petersburg. The object consists of 800 names of restaurants and cafes in the city of St. Petersburg, selected through a continuous sampling method from open internet sources. The purpose of the study is to identify the connotative meanings in the names of restaurants in St. Petersburg from the perspective of linguoculturology. The research is aimed at analyzing how ergonyms reflect the cultural, historical, and social aspects of the city, as well as their role in shaping identities in the context of globalization. The material for the study consisted of the names of 800 restaurants and cafes in St. Petersburg, collected from internet resources. The following methods were applied: descriptive method, continuous sampling method for collecting ergonyms, classificatory analysis to highlight 5 categories of connotative meanings, quantitative counts, and linguocultural interpretation of the connotative meanings of ergonyms. The scientific novelty lies in the systematic analysis of the connotative meanings of ergonyms as a key element of the linguistic picture of the city of St. Petersburg, revealing the connection between nomination, cultural memory, and contemporary sociocultural processes. For the first time, a detailed interpretation of the connotative meanings of ergonyms in St. Petersburg was conducted across five categories. The results showed that the largest category was "Globalization and Intercultural Synthesis" (289 instances), reflecting the adaptation of borrowed elements to the local culture. The category "Gastronomic Pragmatics and Specificity" (183 instances) minimizes cognitive load. The categories "Historical and Cultural References," "Contemporary Cultural Trends and Social Identity," and "Natural Images and Regional Features" are represented in similar numbers (119, 116, and 93 instances respectively). The study confirms that the ergonyms of St. Petersburg shape the "linguistic map" of the city, activating historical, cultural, natural, and gastronomic aspects.