Semantic Coherence Research Articles

Validation of natural language processing methods capturing semantic incoherence in the speech of patients with non-affective psychosis.

Impairments in speech production are a core symptom of non-affective psychosis (NAP). While traditional clinical ratings of patients' speech involve a subjective human factor, modern methods of natural language processing (NLP) promise an automatic and objective way of analyzing patients' speech. This study aimed to validate NLP methods for analyzing speech production in NAP patients. Speech samples from patients with a diagnosis of schizophrenia or schizoaffective disorder were obtained at two measurement points, 6 months apart. Out of N = 71 patients at T1, speech samples were also available for N = 54 patients at T2. Global and local models of semantic coherence as well as different word embeddings (word2vec vs. GloVe) were applied to the transcribed speech samples. They were tested and compared regarding their correlation with clinical ratings and external criteria from cross-sectional and longitudinal measurements. Results did not show differences for global vs. local coherence models and found more significant correlations between word2vec models and clinically relevant outcome variables than for GloVe models. Exploratory analysis of longitudinal data did not yield significant correlation with coherence scores. These results indicate that natural language processing methods need to be critically validated in more studies and carefully selected before clinical application.

Neural Rendering-Based 3D Scene Style Transfer Method via Semantic Understanding Using a Single Style Image

In the rapidly emerging era of untact (“contact-free”) technologies, the requirement for three-dimensional (3D) virtual environments utilized in virtual reality (VR)/augmented reality (AR) and the metaverse has seen significant growth, owing to their extensive application across various domains. Current research focuses on the automatic transfer of the style of rendering images within a 3D virtual environment using artificial intelligence, which aims to minimize human intervention. However, the prevalent studies on rendering-based 3D environment-style transfers have certain inherent limitations. First, the training of a style transfer network dedicated to 3D virtual environments demands considerable style image data. These data must align with viewpoints that closely resemble those of the virtual environment. Second, there was noticeable inconsistency within the 3D structures. Predominant studies often neglect 3D scene geometry information instead of relying solely on 2D input image features. Finally, style adaptation fails to accommodate the unique characteristics inherent in each object. To address these issues, we propose a novel approach: a neural rendering-based 3D scene-style conversion technique. This methodology employs semantic nearest-neighbor feature matching, thereby facilitating the transfer of style within a 3D scene while considering the distinctive characteristics of each object, even when employing a single style image. The neural radiance field enables the network to comprehend the geometric information of a 3D scene in relation to its viewpoint. Subsequently, it transfers style features by employing the unique features of a single style image via semantic nearest-neighbor feature matching. In an empirical context, our proposed semantic 3D scene style transfer method was applied to 3D scene style transfers for both interior and exterior environments. This application utilizes the replica, 3DFront, and Tanks and Temples datasets for testing. The results illustrate that the proposed methodology surpasses existing style transfer techniques in terms of maintaining 3D viewpoint consistency, style uniformity, and semantic coherence.

Retracted: Analysis of Syntactic Complexity and Semantic Coherence of Academic English Writing Based on Particle Swarm Optimization

Retracted: Analysis of Syntactic Complexity and Semantic Coherence of Academic English Writing Based on Particle Swarm Optimization

Expectation-based gist facilitation: Rapid scene understanding and the role of top-down information.

Scene meaning is processed rapidly, with "gist" extracted even when presentation duration spans a few dozen milliseconds. This has led some to suggest a primacy of bottom-up information. However, gist research has typically relied on showing successions of unrelated scene images, contrary to our everyday experience in which the world unfolds around us in a predictable manner. Thus, we investigated whether top-down information-in the form of observers' predictions of an upcoming scene-facilitates gist processing. Within each trial, participants (N = 370) experienced a series of images, organized to represent an approach to a destination (e.g., walking down a sidewalk), followed by a target scene either congruous or incongruous with the expected destination (e.g., a store interior or a bedroom). A series of behavioral experiments revealed that appropriate expectations facilitated gist processing; inappropriate expectations interfered with gist processing; sequentially-arranged scene images benefitted gist processing when semantically related to the target scene; expectation-based facilitation was most apparent when presentation duration was most curtailed; and findings were not simply the result of response bias. We then investigated the neural correlates of predictability on scene processing using event-related potentials (ERPs) (N = 24). Congruency-related differences were found in a putative scene-selective ERP component, related to integrating visual properties (P2), and in later components related to contextual integration including semantic and syntactic coherence (N400 and P600, respectively). Together, results suggest that in real-world situations, top-down predictions of an upcoming scene influence even the earliest stages of its processing, affecting both the integration of visual properties and meaning. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Coh-Metrix 시스템을 활용한 고등학교 영어 모의고사 읽기 지문의 학년 간 연계성 분석

The main objective of this study was to analyze the continuity of reading passages in high school English mock College Scholastic Ability Test (CAST) exams using Coh-Metrix, a multi-level text analysis tool. To achieve this, the study constructed a corpus of reading passages from the 2022 high school English mock exams and subjected them to a broad range of Coh-Metrix measures. These measures included basic measures (the number of words, the number of sentences, average word length, average sentence length), word frequencies (word frequencies for content words), word features (imageability, concreteness, age of acquisition, familiarity), lexical diversity measures (type-token ratios for content words), personal pronouns (first person pronouns, second person pronouns, third person pronouns), connectives, readability indices (Flesch Reading Ease, Flesch-Kincaid Grade Level), syntactic complexity (noun density scores, the number of words before main verbs), coreference cohesion measures, and semantic cohesion measures. The main results revealed that the continuity of high school English mock tests was well established for average word and sentence length measures, the average number of words, word frequencies, word familiarity, second person pronouns, standard readability indices, and syntactic complexity measures. These findings have implications for the development of reading passages in high school English mock exams.

Reconstruction of Semantic Coherence of Characters’ Names from the Perspective of the Translator’s Subjectivity

Reconstruction of Semantic Coherence of Characters’ Names from the Perspective of the Translator’s Subjectivity

VLTinT: Visual-Linguistic Transformer-in-Transformer for Coherent Video Paragraph Captioning

Video Paragraph Captioning aims to generate a multi-sentence description of an untrimmed video with multiple temporal event locations in a coherent storytelling. Following the human perception process, where the scene is effectively understood by decomposing it into visual (e.g. human, animal) and non-visual components (e.g. action, relations) under the mutual influence of vision and language, we first propose a visual-linguistic (VL) feature. In the proposed VL feature, the scene is modeled by three modalities including (i) a global visual environment; (ii) local visual main agents; (iii) linguistic scene elements. We then introduce an autoregressive Transformer-in-Transformer (TinT) to simultaneously capture the semantic coherence of intra- and inter-event contents within a video. Finally, we present a new VL contrastive loss function to guarantee the learnt embedding features are consistent with the captions semantics. Comprehensive experiments and extensive ablation studies on the ActivityNet Captions and YouCookII datasets show that the proposed Visual-Linguistic Transformer-in-Transform (VLTinT) outperforms previous state-of-the-art methods in terms of accuracy and diversity. The source code is made publicly available at: https://github.com/UARK-AICV/VLTinT.

Comparing global and local semantic coherence of spontaneous speech in persons with Alzheimer's disease and healthy controls

There is growing evidence that corpus-based computational tools are useful in identifying changes in speech that appear to accompany cognitive decline in Alzheimer's disease (AD). It has long been known that semantic coherence of speech is altered in AD, but only recently have computational tools been developed that allow for cohesion indices to be computed in an automated fashion on larger data sets. To that end, this study examined semantic coherence in persons with AD and healthy controls. Speech transcripts from 81 individuals with probable AD (Mage = 72.7 years, SD = 8.80, 70.4% female) and 61 healthy controls (Mage = 63.9 years, SD = 8.52, 62.3% female) from DementiaBank were analyzed. Machine learning analyses of coherence were conducted, and models evaluated for classification accuracy (i.e., AD vs controls) as well as ROC-AUC. Relationships between coherence indices and MMSE performance were also quantified. Though no significant group differences emerged in local semantic coherence among adjacent words, persons with AD produced less globally coherent speech relative to healthy controls. Furthermore, global coherence indices predicted AD diagnoses with accuracy between 75% and 78% and were significantly associated with MMSE scores. These findings suggest that automated measures of global coherence can distinguish individuals with AD from healthy controls, which may point to eventual diagnostic utility in clinical settings.

“Such Common Objects” – The Politics of Everyday Life in G. K. Chesterton’s and Rose Macaulay’s Thing-Essays

In this article I argue that so-called light essays on apparently trivial items as they were published in periodicals of the early 20th century use the specificities of the essay genre to unveil dimensions of everyday life that are often taken for granted and thus overlooked. I will focus on two thing-essays – G. K. Chesterton’s “Lamp-Posts” (1920) and Rose Macaulay’s “Arm-Chair” (1935) – and argue that these essays help an emerging mass reader- and consumership to establish semantic coherence between themselves and the proliferating objects around them. Using Martin Heidegger’s and Bill Brown’s object-thing-distinction, I will also argue that these essays point to the relevance of everyday life and to the political issues that are sometimes manifest in everyday items. In particular, Chesterton’s and Macaulay’s thing-essays use the semantic flexibility of things to illustrate and defend their political, religious and intellectual views, touching on fields such as distributism, Catholicism and the so-called battle of the brows.

A knowledge graph-based deep learning framework for efficient content similarity search of Sustainable Development Goals data

Abstract Sustainable development denotes the enhancement of living standards in the present without compromising future generations’ resources. Sustainable Development Goals (SDGs) quantify the accomplishment of sustainable development and pave the way for a world worth living in for future generations. Scholars can contribute to the achievement of the SDGs by guiding the actions of practitioners based on the analysis of SDG data, as intended by this work. We propose a framework of algorithms based on dimensionality reduction methods with the use of Hilbert Space Filling Curves (HSFCs) in order to semantically cluster new uncategorised SDG data and novel indicators, and efficiently place them in the environment of a distributed knowledge graph store. First, a framework of algorithms for insertion of new indicators and projection on the HSFC curve based on their transformer-based similarity assessment, for retrieval of indicators and load-balancing along with an approach for data classification of entrant-indicators is described. Then, a thorough case study in a distributed knowledge graph environment experimentally evaluates our framework. The results are presented and discussed in light of theory along with the actual impact that can have for practitioners analysing SDG data, including intergovernmental organizations, government agencies and social welfare organizations. Our approach empowers SDG knowledge graphs for causal analysis, inference, and manifold interpretations of the societal implications of SDG-related actions, as data are accessed in reduced retrieval times. It facilitates quicker measurement of influence of users and communities on specific goals and serves for faster distributed knowledge matching, as semantic cohesion of data is preserved.

WETM: A word embedding-based topic model with modified collapsed Gibbs sampling for short text

Short texts are a common source of knowledge, and the extraction of such valuable information is beneficial for several purposes. Traditional topic models are incapable of analyzing the internal structural information of topics. They are mostly based on the co-occurrence of words at the document level and are often unable to extract semantically relevant topics from short text datasets due to their limited length. Although some traditional topic models are sensitive to word order due to the strong sparsity of data, they do not perform well on short texts. In this paper, we propose a novel word embedding-based topic model (WETM) for short text documents to discover the structural information of topics and words and eliminate the sparsity problem. Moreover, a modified collapsed Gibbs sampling algorithm is proposed to strengthen the semantic coherence of topics in short texts. WETM extracts semantically coherent topics from short texts and finds relationships between words. Extensive experimental results on two real-world datasets show that WETM achieves better topic quality, topic coherence, classification, and clustering results. WETM also requires less execution time compared to traditional topic models.

Attention guided domain alignment for conditional face image generation

Despite the great success of Generative Adversarial Network (GAN) in face image generation, it is still a challenge to generate a faithful yet high-fidelity face image given an exemplar image and a conditional input from a distinct domain (e.g., a semantic segmentation mask or an edge map). Existing methods learn image-level features to align distinct domains in an intermediate domain, which ignore the spatial relationship of the facial semantic parts and therefore lead to semantic mismatching problems. In addition, it is computationally expensive to establish dense correspondences in the intermediate domain, especially for high-resolution face images. To address these problems, we propose a novel attention guided domain alignment method for conditional face image generation, which aligns two domains directly under the guidance of the local attention learned from semantically similar face parts. In particular, we assign a dedicated index for each feature block and adopt a top-k ranking operation to correspond block-wise features in two distinct domains, which exploits the spatial relationship of facial parts and preserves the texture structure during alignment. The local attention is then learned from the retrieved blocks, which reduces computation complexity substantially and therefore enables to build alignment in a high resolution. The aligned features are finally fused with adaptive weights learned from their long-range correlation coefficients, which capture the semantic coherence of the style features between the two domains. Extensive experimental results on the CelebAMask-HQ dataset demonstrate that the proposed method is superior to the state-of-the-art methods.

Neural Image Parts Group Search for Person Re-Identification

Employing partition strategy to explore fine-grained features has been verified to be beneficial for person re-identification in recent literature. However, existing methods primarily rely on expert experience to manually design various partition strategies, which may lead to a sub-optimal solution for fine-grained features exploration. In this paper, we propose a Neural Parts Group Search (NPGS) strategy that auto-searches the optimal parts group via evolutionary algorithm (EA) to facilitate the network to exploit the local details. And during search process, designing a high-quality search space is especially crucial for an efficient optimization. Considering the human top-down structure and the semantic coherence of parts, we design a coarse-to-fine parts search space (C2F-PSP) in NPGS, which effectively reduce the search complexity without the loss of parts expressivity. Additionally, since only employing the high-level semantic features is insufficient for the NPGS to search effective parts, we further develop an efficient feature aggregation strategy named hierarchical low-rank bilinear pooling that progressively integrates the high-level semantic property and the low-level fine-grained details to facilitate the NPGS to explore the fine-grained features. Furthermore, to relieve the interference of background during parts search process, we propose a novel Relational Attention Module (RAM) by exploiting the channel and spatial structural interdependence of pixels to strengthen the discriminative regions. Extensive experiments on the mainstream evaluation datasets demonstrate that our method outperforms the recent state-of-the-art Re-ID models.

Automatic generation of paragraph templates using a Textual Energy approach:

In this paper, we present the results of preliminary experiments using the Textual Energy measure to be used in Automatic Text Generation tasks (ATG). Textual Energy calculates the similarity among the sentences of a document, using intuitive ideas coming from Mechanical statistical like the associative memories and the energy of a system. Using this approach we intend to generate a set of selected sentences having a semantic and structural coherence. In our experiment, the number of selected sentences was manually determined. In particular, the experiments were performed using sets of 4 sentences. Then, the selected sentences can be employed for paragraph generation using Canned Text-like techniques. We have performed an important number of experiments, and we found interesting results that we present in this paper. These results allow us to conclude that it is possible to generate a set of sentences, as paragraphs-like, through methods, avoiding as much as possible undesirable phenomena, such hallucination, which have been recently found in ATG models based on Deep learning Neural Networks.

English Translation proofreading System based on Information Technology: Construction of semantic Ontology Translation Model

ABSTRACT Translation as a bridge between different cultures has received significant attention in recent years, especially with the rise of globalization and the increasing communication between nations. Translation has become a crucial tool for connecting people, promoting cross-cultural understanding, and facilitating international trade, academic exchange, and outbound tourism. With the advancement of information technology, research on improving English translation algorithms has become a hot topic. This is due to the complexities and variability of language, where words can have different meanings in different contexts. To address the issue of semantic coherence in current English translation proofreading systems, this paper proposes a semantic ontology translation model that utilizes fuzzy mapping to translate English and selects the most semantically appropriate translation using a decision function. The contributions of this research are twofold. Firstly, it provides a new approach to English translation proofreading by incorporating semantic ontology and fuzzy mapping. Secondly, it demonstrates the effectiveness of this approach through the improved efficiency and accuracy of English proofreading. The key findings of this study show that the proposed semantic ontology translation model can significantly improve the semantic coherence of English translation proofreading. This makes this article different from other studies by offering a unique solution to the problem of semantic coherence in English translation proofreading.

A Narrative Review of Speech and EEG Features for Schizophrenia Detection: Progress and Challenges.

Schizophrenia is a mental illness that affects an estimated 21 million people worldwide. The literature establishes that electroencephalography (EEG) is a well-implemented means of studying and diagnosing mental disorders. However, it is known that speech and language provide unique and essential information about human thought. Semantic and emotional content, semantic coherence, syntactic structure, and complexity can thus be combined in a machine learning process to detect schizophrenia. Several studies show that early identification is crucial to prevent the onset of illness or mitigate possible complications. Therefore, it is necessary to identify disease-specific biomarkers for an early diagnosis support system. This work contributes to improving our knowledge about schizophrenia and the features that can identify this mental illness via speech and EEG. The emotional state is a specific characteristic of schizophrenia that can be identified with speech emotion analysis. The most used features of speech found in the literature review are fundamental frequency (F0), intensity/loudness (I), frequency formants (F1, F2, and F3), Mel-frequency cepstral coefficients (MFCC's), the duration of pauses and sentences (SD), and the duration of silence between words. Combining at least two feature categories achieved high accuracy in the schizophrenia classification. Prosodic and spectral or temporal features achieved the highest accuracy. The work with higher accuracy used the prosodic and spectral features QEVA, SDVV, and SSDL, which were derived from the F0 and spectrogram. The emotional state can be identified with most of the features previously mentioned (F0, I, F1, F2, F3, MFCCs, and SD), linear prediction cepstral coefficients (LPCC), linear spectral features (LSF), and the pause rate. Using the event-related potentials (ERP), the most promissory features found in the literature are mismatch negativity (MMN), P2, P3, P50, N1, and N2. The EEG features with higher accuracy in schizophrenia classification subjects are the nonlinear features, such as Cx, HFD, and Lya.

Selecting the Number and Labels of Topics in Topic Modeling: A Tutorial

Topic modeling is a type of text analysis that identifies clusters of co-occurring words, or latent topics. A challenging step of topic modeling is determining the number of topics to extract. This tutorial describes tools researchers can use to identify the number and labels of topics in topic modeling. First, we outline the procedure for narrowing down a large range of models to a select number of candidate models. This procedure involves comparing the large set on fit metrics, including exclusivity, residuals, variational lower bound, and semantic coherence. Next, we describe the comparison of a small number of models using project goals as a guide and information about topic representative and solution congruence. Finally, we describe tools for labeling topics, including frequent and exclusive words, key examples, and correlations among topics.

A Systemic Functional Linguistic Analysis of Clauses Relationship in Luke Gospel Text, Janji Baru Using Kupang Malay

This paper reports on a functional linguistics study of clauses relationship in which it is included as textual meta function of meaning. This textual metafunction is on the level of lexicogrammar. Semantic sequence of figures is realized by a clause complexes series. This paper focused on relationship types between clauses, interdependency and logico-semantic relationship. Data were taken from A Bible, New Testament using Kupang Malay called Janji Baru, especially in Luke of Gospel text. The analysis revealed that parataxis used in text covered coordinate conjunctions such as, deng ’and’, ma, ‘but’, jadi/ko ‘so, for’, ais ju ‘then’. The hypotaxis used subordinate conjunctions, such as, kalo ‘if’’ tagal ‘because/since’, te ‘cause’. The use of parataxis and hypotaxis indicates that semantic sequence is coherence in which conjunctions used in text are usually repeated. It is the same as the use of logico-semantic, such as expansion and projection. The expansion consisted of elaboration, extension, and enhancement. The projection consisted of locution and idea that related to phenomena of a higher order of experience. The idea used verbal process, such as, kastau bilang ‘tell/say’, manyao bilang, ’answer’, tanya bilang ‘ask’, etc. This fact indicates that the idea stated was stressed by using serial verbs or usually a verb was ended by other verb’ This fact indicates to attract listeners’ attention. The locution used mental process, such as, rasa ‘feel’ ‘inga’ ‘remember’. Thus, it can be stated that the use of interdependency and logico-semantic relation was aimed to keep the semantic sequence and text coherence.

Language Analytics for Assessment of Mental Health Status and Functional Competency.

Automated language analysis is becoming an increasingly popular tool in clinical research involving individuals with mental health disorders. Previous work has largely focused on using high-dimensional language features to develop diagnostic and prognostic models, but less work has been done to use linguistic output to assess downstream functional outcomes, which is critically important for clinical care. In this work, we study the relationship between automated language composites and clinical variables that characterize mental health status and functional competency using predictive modeling. Conversational transcripts were collected from a social skills assessment of individuals with schizophrenia (n = 141), bipolar disorder (n = 140), and healthy controls (n = 22). A set of composite language features based on a theoretical framework of speech production were extracted from each transcript and predictive models were trained. The prediction targets included clinical variables for assessment of mental health status and social and functional competency. All models were validated on a held-out test sample not accessible to the model designer. Our models predicted the neurocognitive composite with Pearson correlation PCC = 0.674; PANSS-positive with PCC = 0.509; PANSS-negative with PCC = 0.767; social skills composite with PCC = 0.785; functional competency composite with PCC = 0.616. Language features related to volition, affect, semantic coherence, appropriateness of response, and lexical diversity were useful for prediction of clinical variables. Language samples provide useful information for the prediction of a variety of clinical variables that characterize mental health status and functional competency.

It is difficult to think about complex things (tools of new pedagogical thinking)

A new level of systemic complexity of the tasks facing modern education determines the need to search for new thinking tools that ensure their solution. This is the reason for the relevance of this research, the purpose of which is to offer a new thinking tool that provides the opportunity to think complexly and about complex things, to view an object, for example, education, as a complex whole, i.e. included through the generality of the constitutive basis in a system of diverse connections and interactions$ Qualitatively compress large arrays of heterogeneous information, integrate (collect) it into semantic coherence. In the process of scientific research, methods of scientific theorization were used in the context of the ideas of post-nonclassical science and on its methodological basis. The article considers the structure of the ontological representation as a tool of thinking, which is understood as the order of division of the world in the consciousness of society (how the meaning of this representation organizes consciousness). This is the scientific novelty of the study. Using the example of a traditional learning system, the fundamental influence of the structure of the ontological representation in which it is built (the separation between the subjective world and the objective world) is shown. From this position, four structural markers of this system and its basic meanings (ontological, anthropological, etc.) are identified. Their presence in educational (in) innovations indicates that these (in) innovations are a variant of the traditional learning system, respectively, the quality of the educational result and its inconsistency with the challenges of the time are common. The analysis of FGOS 4.0 — the first digital standard confirms this. The article examines the structure of the ontological foundation of a new education system capable of responding to the challenges of the time, its basic meanings are highlighted. It is concluded that in the new system, along with new educational opportunities, the educational opportunities of the traditional system of education should be used.