Semantic Concepts Research Articles

SaGCN: Semantic-Aware Graph Calibration Network for Temporal Sentence Grounding

Temporal sentence grounding is a challenging task that aims to localize the semantic corresponding segment from the untrimmed video according to the given query language description. Existing methods either utilize a cross-modal matching architecture following a scan-and-rank pipeline or directly predict the probabilities of being the target boundary for each frame based on the entire video content. However, such methods are weak when some of the critical semantic concepts in the query are actually relevant to multiple video segments or the desired video segment contains a query-irrelevant scene due to ignoring query semantic concepts and local and global cross-modal context. In this paper, we propose a novel semantic-aware graph calibration network (SaGCN) to address the issues mentioned above. Specifically, we first introduce a semantic-aware local relational graph module to capture the inherent relationships among the specific semantic concept relevant local contextual information for fine-grained cross-modal information interactions. Then, a semantic-aware global relational graph module is derived for global contextual information integration and achieving cross-modal alignment. Finally, an attention-based calibration module is designed for eliminating the irrelevant information maintained in the visual modality under the guidance of query description. Extensive experiments verify the effectiveness of our proposed SaGCN on two widely used datasets (Charades-STA and TACoS), in which we achieve significant and consistent improvement compared to the state-of-the-art approaches.

Semantics-Native Communication via Contextual Reasoning

Recently, machine learning (ML) has shown its effectiveness in improving communication efficiency by reinstating the semantics of bits. To understand its underlying principles, we propose a novel stochastic model of semantic communication, dubbed semantics-native communication (SNC). Inspired from human communication, we consider a point-topoint SNC scenario where a speaker has an intention of referring to an entity, extracts its semantic concepts, and maps these concepts into symbols that are communicated to a target listener through the traditional Shannon communication channel. Next, we recall rational humans who can communicate more efficiently by reasoning about the others’ contexts before communication, referred to as contextual reasoning. This motivates us to propose a novel SNC framework harnessing agent states as side information in a way that the speaker locally and iteratively communicates with a virtual agent having the listener’s state, and vice versa. Theoretically, we prove the convergence of contextual reasoning, at which it minimizes the bit length while guaranteeing a target reliability. Simulation results corroborate that contextual reasoning based SNC can significantly reduce bit lengths, and be a robust solution to imperfect agent states by quantizing the entity-concept-symbol mapping before contextual reasoning.

Improving semantic coverage of data-to-text generation model using dynamic memory networks

AbstractThis paper proposes a sequence-to-sequence model for data-to-text generation, called DM-NLG, to generate a natural language text from structured nonlinguistic input. Specifically, by adding a dynamic memory module to the attention-based sequence-to-sequence model, it can store the information that leads to generate previous output words and use it to generate the next word. In this way, the decoder part of the model is aware of all previous decisions, and as a result, the generation of duplicate words or incomplete semantic concepts is prevented. To improve the generated sentences quality by the DM-NLG decoder, a postprocessing step is performed using the pretrained language models. To prove the effectiveness of the DM-NLG model, we performed experiments on five different datasets and observed that our proposed model is able to reduce the slot error rate rate by 50% and improve the BLEU by 10%, compared to the state-of-the-art models.

Book as an Artistic Artifact

A book is not only a carrier of information, but also the way of its aesthetic visualization. This actualizes the problem of understanding a book as an artistic artifact. The study of the book in this context implies the identification of several aspects: the art of the book; the existence of the book as an exhibit of museum and exhibition art space; the use of the book as a compositional and meaningful element of works of fine art, as a component of library design, living and working spaces; alterbooking and architectural structures in the form of a book. The article is devoted to the art of the book and its exhibition in museum and exhibition space. Its goal is to systematize the factors and features that should be considered in the study of the art of book for the possibility of its holistic interpretation from an art criticism perspective. The article considers the role of decoration in creating visual and verbal unity. The functions of illustration depending on the cultural and historical context as well as a purpose of the book are determined. Attention is paid to such phenomena of book art as fore-edge painting, artist’s book and art albums. The existence of the book in museum and exhibition art space is also considered. There are two types of exhibitions: single-subject exhibitions devoted exclusively to books and exhibitions with other objects participating in the realization of the semantic concept of a cultural event.

Unsupervised Hashing with Semantic Concept Mining

Recently, to improve the unsupervised image retrieval performance, plenty of unsupervised hashing methods have been proposed by designing a semantic similarity matrix, which is based on the similarities between image features extracted by a pre-trained CNN model. However, most of these methods tend to ignore high-level abstract semantic concepts contained in images. Intuitively, concepts play an important role in calculating the similarity among images. In real-world scenarios, each image is associated with some concepts, and the similarity between two images will be larger if they share more identical concepts. Inspired by the above intuition, in this work, we propose a novel Unsupervised Hashing with Semantic Concept Mining, called UHSCM, which leverages a VLP model to construct a high-quality similarity matrix. Specifically, a set of randomly chosen concepts is first collected. Then, by employing a vision-language pretraining (VLP) model with the prompt engineering which has shown strong power in visual representation learning, the set of concepts is denoised according to the training images. Next, the proposed method UHSCM applies the VLP model with prompting again to mine the concept distribution of each image and construct a high-quality semantic similarity matrix based on the mined concept distributions. Finally, with the semantic similarity matrix as guiding information, a novel hashing loss with a modified contrastive loss based regularization item is proposed to optimize the hashing network. Extensive experiments on three benchmark datasets show that the proposed method outperforms the state-of-the-art baselines in the image retrieval task.

Interactive and incremental learning of spatial object relations from human demonstrations.

Humans use semantic concepts such as spatial relations between objects to describe scenes and communicate tasks such as "Put the tea to the right of the cup" or "Move the plate between the fork and the spoon." Just as children, assistive robots must be able to learn the sub-symbolic meaning of such concepts from human demonstrations and instructions. We address the problem of incrementally learning geometric models of spatial relations from few demonstrations collected online during interaction with a human. Such models enable a robot to manipulate objects in order to fulfill desired spatial relations specified by verbal instructions. At the start, we assume the robot has no geometric model of spatial relations. Given a task as above, the robot requests the user to demonstrate the task once in order to create a model from a single demonstration, leveraging cylindrical probability distribution as generative representation of spatial relations. We show how this model can be updated incrementally with each new demonstration without access to past examples in a sample-efficient way using incremental maximum likelihood estimation, and demonstrate the approach on a real humanoid robot.

METAPHORICAL FRAMINGS IN THE NEW YORK TIMES ONLINE PRESS REPORTS ABOUT CHATGPT

Through a combined quantitative-qualitative approach the paper aims to explore the densities and possible rhetorical effects of the most common metaphorical framings used in the New York Times online press reports about ChatGPT. The theoretical framework is based on the main tenets of conceptual metaphor theory (Lakoff and Johnson 2003[1980]), frame semantics (Fillmore 1982), and image schema theory (Johnson 1987). The initial stage involved the construction of a small specialized corpus which was tagged manually for the analysis in WordSmith Tools 6.0. Metaphor identification was conducted in line with MIPVU (STEEN et al. 2010). Quantitative analysis showed the highest density for containment, motion, and force metaphors. Living being, tool, conflict, and machine metaphors were also present, but not as frequent. The linear regression model (motion, force, living being) explained 61% of variance in the density of containment metaphors, and it was significant (p=.002). Additional linear regression models have also been tested. The results suggest the potential for interaction of these metaphor groups when they cooccur in discourse (i.e., they may work in concert to construct a specific viewpoint). Qualitative analysis showed a range of possible rhetorical effects that the identified metaphorical framings may play (e.g., making AI more appealing by framing it as a living being or tool, raising doubts about the use of AI by constructing borders and containers, or framing AI as a threat). Results of qualitative analysis should be understood as testable hypotheses that remain to be further explored and validated in experimental settings.

Trump’s Election Billboards: A Multimodal Cognitive Study

Multimodality is the analysis of language besides other modalities (images, sounds, gestures, video images, etc.). Election billboards as a kind of multimodal text is used during elections to promote a candidate's image and advertise their presence. The current study seeks to analyze a sample of Trump’s 2020 election billboards in the American presidential election billboards as a type of multimodal texts from the view point of Multimodality and Cognitive Semantics. Kress and van Leeuwen's (2006) model and Langcker’s Construal Theory (2013) are used to demystify how the visual and linguistic elements cooperate to serve the purpose of the billboard as a whole. The visual elements are analyzed based on Kress and van Leeuwen (2006) Grammar of Visual Design model. On the other hand, Langacker's construal operation model is used for the analysis of both linguistic and visual elements as a cognitive semantic theory. The analysis's findings indicate that the models are essential to constructing and deciphering the meaning of political election campaign billboards, and they also prove that in constructing meaning, the visual and textual elements of Trump’s election billboards complement one another even though the focus on the linguistic elements is observed to be higher to some extent rather than that of the visual elements

BIM product recommendation for intelligent design using style learning

Building information modeling (BIM) has been widely adopted in architectural interior design due to its characteristics such as visualization, coordination and simulation. Currently, an increasing number of BIM products are created to facilitate BIM interior design, which triggers new demand for BIM product recommendations to improve design efficiency by directly recommending satisfactory BIM products for designers. Current efforts mainly focus on the text-based BIM product retrieval and ignore the high-level semantic concept of style features. Although style consistency is one of the critical design principles, BIM product recommendation satisfying style consistency remains to be unexplored. This study proposes a novel design recommendation scheme for BIM products using style learning, termed DrStyle. Firstly, a product representative image selection model is designed to identify a representative image retaining the best style of a BIM product. Then, the DrStyle employs a style learning algorithm to extract the style feature vector using the representative image as input. Finally, a BIM product recommendation scheme is proposed to recommend BIM products by computing the similarities of style feature vectors. The performance of the proposed DrStyle was evaluated, and the experiment results showed that the DrStyle achieved an average precision of 68.8% purely using the style. This study complements intelligent design from a novel perspective of style consistency and will inspire more comprehensive intelligent design schemes.

Improving Word Embedding Using Variational Dropout

Pre-trained word embeddings are essential in natural language processing (NLP). In recent years, many post-processing algorithms have been proposed to improve the pre-trained word embeddings. We present a novel method - Orthogonal Auto Encoder with Variational Dropout (OAEVD) for improving word embeddings based on orthogonal autoencoders and variational dropout. Specifically, the orthogonality constraint encourages more diversity in the latent space and increases semantic similarities between similar words, and variational dropout makes it more robust to overfitting. Empirical evaluation on a range of downstream NLP tasks, including semantic similarity, text classification, and concept categorization shows that our proposed method effectively improves the quality of pre-trained word embeddings. Moreover, the proposed method successfully reduces the dimensionality of pre-trained word embeddings while maintaining high performance.

A Cognitive Semantics Analysis of the Directional Motion Verbs wang往 and lai 来in the YiChing

The Chinese directional motion verbs wang往 and lai来can be glossed in English as ‘go’ and ‘come’, and the distinction between ‘go’ and ‘come’ in English is comparable to that between wang and lai in Chinese. Their usage in the YiChing is described and analyzed in this article. In the YiChing, wang and lai frequently co-occur in the same sentence, as well as occurring singly elsewhere. Based on a corpus analysis of the YiChing, combined with qualitative interpretive analysis, this article proposes that the majority of co-occurring usages of wang and lai are metaphorical. Metaphorical uses can also be found in the single usages of the two verbs. These metaphors are informed by the cultural practices and values that form the context for the YiChing as a cultural artifact as well as a classical text. Further, these directional motion verbs are based on metaphorical extensions that lead to their semantic changes (metaphors are based on the motion). Thus, the article further analyzes the role of metaphor and cultural context in the semantic changes to the directional motion verbs wang and lai.

Torsion Geometry 5-Fold Symmetry, Anholonomic Phases, Klein Bottle Logophysics, Chaos, Resonance: Applications Towards a Novel Paradigm for the Neurosciences and Consciousness

Torsion Geometry 5-Fold Symmetry, Anholonomic Phases, Klein Bottle Logophysics, Chaos, Resonance: Applications Towards a Novel Paradigm for the Neurosciences and Consciousness

ON THE STUDY OF THE CONCEPT TERM IN WORLD LINGUISTICS

In this article, the study of such terms as concept, conceptualization, categorization, conceptual semantics, linguistic semantics, microconcept, macroconcept, verbalizer, verbal means, non-verbal means, which are widely used in world linguistics, by linguists, as well as the opinions put forward by each scientist within the framework of terms are being studied. It is also explained that when a person understands, compares, compares and observes the realities of material existence, the term concept and the process of categorization are comprehended, a conceptual picture of the world composed of concepts is formed in his mind, and the expression of these concepts through the language units of certain language levels.

Resolving Cross-modal Semantic Interference among Object Concepts Requires Medial Temporal Lobe Cortex

The ability to flexibly categorize object concepts is essential to semantic cognition because the features that make two objects similar in one context may be irrelevant and even constitute interference in another. Thus, adaptive behavior in complex and dynamic environments requires the resolution of feature-based interference. In the current case study, we placed visual and functional semantic features in opposition across object concepts in two categorization tasks. Successful performance required the resolution of functional interference in a visual categorization task and the resolution of visual interference in a functional categorization task. In Experiment 1, we found that patient D. A., an individual with bilateral temporal lobe lesions, was unable to categorize object concepts in a context-dependent manner. His impairment was characterized by an increased tendency to incorrectly group objects that were similar on the task-irrelevant dimension, revealing an inability to resolve cross-modal semantic interference. In Experiment 2, D. A.'s categorization accuracy was comparable to controls when lures were removed, indicating that his impairment is unique to contexts that involve cross-modal interference. In Experiment 3, he again performed as well as controls when categorizing simple concepts, suggesting that his impairment is specific to categorization of complex object concepts. These results advance our understanding of the anterior temporal lobe as a system that represents object concepts in a manner that enables flexible semantic cognition. Specifically, they reveal a dissociation between semantic representations that contribute to the resolution of cross-modal interference and those that contribute to the resolution of interference within a given modality.

Color as the meaning-forming concept of Zhang Yimou's film "Hero"

The subject of the study is the color meaning in the film of the famous Chinese director Zhang Yimou "Hero", the conceptual understanding of color as the main semantic element of the artistic solution. The object of research is the techniques of visual solutions in cinema. The perception of color in the culture of different peoples differs in certain nuances and depends on many factors, including national traditions and historically established stereotypes. Films by Chinese, Korean, and Japanese filmmakers are interesting from the point of view of color solutions, as they differ in artistic originality and often offer the Russian audience an extraordinary semantic solution of the idea not only at the level of comprehension of the content, but also at the level of a specific color palette. The author similarly examines the color symbolism of Zhang Yimou's film "Hero". The conceptual content of the visual-figurative series, the originality of the coloristic solution of the film is comprehended. The scientific novelty of the article lies in the study of the color symbolism of Zhang Yimou's film "Hero" in the aspect of the correlation between the perception of the coloristic solution of the visual series and the conceptual content of the film. The author comes to the conclusion that despite the ambivalence of perception of a particular color in visual art, the coloristic solution is predetermined by the conceptual content of each episode, so color becomes the leading semantic concept of Zhang Yimou's film "Hero".

Davidson on Truth

"Truth-conditional semantics is by far the best-known philosophical contribution of Donald Davidson. The main idea of this approach is to explain the concept of meaning by appeal to the concept of truth. Accordingly, we understand a sentence s of a natural language L, if and only if, we know its truth-conditions. Challenging in its nature, this proposal immediately caught the attention of the philosophical community, being equally appreciated and criticized. The aim of this paper is to argue that Davidson’s approach is too optimistic in its goals. In my view, truth-conditional semantics is unlikely to be the best way to shed light on the concept of meaning. By appealing to Tarski’s semantic conception of truth and assuming a primitive concept of truth, this perspective leaves too many questions unanswered and thus proves its limits. Keywords: Donald Davidson, meaning, truth, Truth-Conditional Semantics. "

Users' Reactions to Announced Vaccines Against COVID-19 Before Marketing in France: Analysis of Twitter Posts.

Within a few months, the COVID-19 pandemic had spread to many countries and had been a real challenge for health systems all around the world. This unprecedented crisis has led to a surge of online discussions about potential cures for the disease. Among them, vaccines have been at the heart of the debates and have faced lack of confidence before marketing in France. This study aims to identify and investigate the opinions of French Twitter users on the announced vaccines against COVID-19 through sentiment analysis. This study was conducted in 2 phases. First, we filtered a collection of tweets related to COVID-19 available on Twitter from February 2020 to August 2020 with a set of keywords associated with vaccine mistrust using word embeddings. Second, we performed sentiment analysis using deep learning to identify the characteristics of vaccine mistrust. The model was trained on a hand-labeled subset of 4548 tweets. A set of 69 relevant keywords were identified as the semantic concept of the word "vaccin" (vaccine in French) and focused mainly on conspiracies, pharmaceutical companies, and alternative treatments. Those keywords enabled us to extract nearly 350,000 tweets in French. The sentiment analysis model achieved 0.75 accuracy. The model then predicted 16% of positive tweets, 41% of negative tweets, and 43% of neutral tweets. This allowed us to explore the semantic concepts of positive and negative tweets and to plot the trends of each sentiment. The main negative rhetoric identified from users' tweets was that vaccines are perceived as having a political purpose and that COVID-19 is a commercial argument for the pharmaceutical companies. Twitter might be a useful tool to investigate the arguments for vaccine mistrust because it unveils political criticism contrasting with the usual concerns on adverse drug reactions. As the opposition rhetoric is more consistent and more widely spread than the positive rhetoric, we believe that this research provides effective tools to help health authorities better characterize the risk of vaccine mistrust.

Relations across Cognitive Faculties: An Addition to the Taxonomy of Cognitive Semantics

Abstract Language is often approached as a self-contained system, one with its own specifically linguistic elements of organization, generally independent of other systems in cognition. But by the analysis here, language shares parts of its organization with other systems in cognition and could not function without their participation. For this analysis, cognition is heuristically divided into a number of cognitive faculties, each judged to perform some integrated function. Some faculties, including language, are treated as “cognitive systems” and others as “cognitive organizers”. Language is examined both in its evolutionary relation to other faculties and as an interface among them.

False triviality of truth

As it stands, Frege’s semantic theory does not provide sufficient general keys for interpreting assertible sentences at different levels of complexity, such as (a) counterfactual-modal-intensional sentences and (b) extensional sentences. It is possible to devise that general key by adding non-classical parameters such as possible worlds, which allow sentences with a high degree of non-extensional complexity to occupy a trivial place in Tarski’s hierarchy as T-scheme-eligible substitutes. The paper argues that finding these non-false (if true) conditions under which complex propositions can be trivially treated as eligible T-scheme substitutes is not trivial. It is a challenge that requires several consistency improvements to deal with various competing extensions of the predicate “true”. We will conclude that Frege-Tarski’s semantic conception (enriched by non- classical presuppositions), misrepresents that challenge. It distorts the problem we face in daily practice to strengthen our assertion systems, plan successful assertion strategies, and protect our assertions from semantic value reversals. The true challenge for us is pragmatic and cannot confuse (1) complicated scientific and empirical conditions of assertiveness with (2) the rewardable scoring-conditions under which one learns to use sentences in his native language.

The temporal mechanism of contextual emotions’ effects on novel abstract and concrete word learning

While previous studies have explored the neurocognitive mechanism of novel word learning under sentential contexts, few studies have probed the potential effects of contextual emotions. The current study used behavioral methods and electroencephalography to elucidate the dynamic influences of contextual emotions while encoding abstract or concrete novel words on subsequent retrieval of the learned meanings. Enhanced N400 and smaller late positive components (LPC) were elicited when encoding abstract words in the positive context. Furthermore, negative contexts were found to hinder both abstract and concrete word learning, while positive contexts had a greater inhibitory effect on learning abstract words than concrete words. Taken together, these findings suggest that the effects of contextual emotions on novel word learning depend on the concreteness of the learned word and occur during the early and late processing stages while mapping abstract and concrete semantic concepts to novel words.