Partial Semantics Research Articles

Future Swedish 3D City Models—Specifications, Test Data, and Evaluation

Three-dimensional city models are increasingly being used for analyses and simulations. To enable such applications, it is necessary to standardise semantically richer city models and, in some cases, to connect the models with external data sources. In this study, we describe the development of a new Swedish specification for 3D city models, denoted as 3CIM, which is a joint effort between the three largest cities in Sweden—Stockholm, Gothenburg, and Malmö. Technically, 3CIM is an extension of the OGC standard CityGML 2.0, implemented as an application domain extension (ADE). The ADE is semantically thin, mainly extending CityGML 2.0 to harmonise with national standards; in contrast, 3CIM is mainly based on linkages to external databases, registers, and operational systems for the semantic part. The current version, 3CIM 1.0, includes various themes, including Bridge, Building, Utility, City Furniture, Transportation, Tunnel, Vegetation, and Water. Three test areas were created with 3CIM data, one in each city. These data were evaluated in several use-cases, including visualisation as well as daylight, noise, and flooding simulations. The conclusion from these use-cases is that the 3CIM data, together with the linked external data sources, allow for the inclusion of the necessary information for the visualisation and simulations, but extract, transform, and load (ETL) processes are required to tailor the input data. The next step is to implement 3CIM within the three cities, which will entail several challenges, as discussed at the end of the paper.

Aspectual Coercion and Lexical Semantics Part 1: Using Selection to Describe the Interaction between Construction and Verb Meaning

Abstract Coercion is an inferential strategy used to resolve conflict between an operator and its argument. Such conflicts are resolved in favor of the semantic requirements of the operator (Talmy, 2000). Jackendoff (1997) and De Swart (1998), among others, represent coercion through type-shifting operators that intervene between an aspectual operator and its situation-type argument, ensuring that the argument is of the appropriate type for the operator. This framework has a mapping problem: the rules that it uses to represent aspectual-type shifts simply replace one aspectual type (the input) with another (the output), so it does not explain how the input representation constrains the output representation. This article offers a solution to the mapping problem: treating aspectual type shifts as operations on the decomposed semantic representations of verbs. I will show that two such operations can capture both implicit and explicit aspectual type-shifts in English, involving both tense constructions and aspectual constructions.

Технология создания цифровой карты сельскохозяйственных угодий на территории сельскохозяйственных организаций Ленинградской области с применением QGIS

The article deals with the technology of creating a digital map of agricultural lands in QGIS in order to form an information base on their state in individual farms. As a result of the conducted research, the authors identified the main stages of creating a digital map, defined key tasks at the stage of spatial reference of raster images and proposed ways to solve them. We developed an algorithm for creating an attribute table using the Python programming language, which enables avoiding errors in the process of filling in the semantic part of the generated information base. The main tools for digitizing raster images are highlighted, and in order to identify mistakes in the process of forming its graphical part, a topological editing module is to be used. The obtained research results can be helpful for the executive authorities of the Russian Federation subjects, local self-government bodies to make decisions in the field of land and property management of a separate entity.

Reformulating Graph Kernels for Self-Supervised Space-Time Correspondence Learning.

Self-supervised space-time correspondence learning utilizing unlabeled videos holds great potential in computer vision. Most existing methods rely on contrastive learning with mining negative samples or adapting reconstruction from the image domain, which requires dense affinity across multiple frames or optical flow constraints. Moreover, video correspondence prediction models need to uncover more inherent properties of the video, such as structural information. In this work, we propose HiGraph+, a sophisticated space-time correspondence framework based on learnable graph kernels. By treating videos as a spatial-temporal graph, the learning objective of HiGraph+ is issued in a self-supervised manner, predicting the unobserved hidden graph via graph kernel methods. First, we learn the structural consistency of sub-graphs in graph-level correspondence learning. Furthermore, we introduce a spatio-temporal hidden graph loss through contrastive learning that facilitates learning temporal coherence across frames of sub-graphs and spatial diversity within the same frame. Therefore, we can predict long-term correspondences and drive the hidden graph to acquire distinct local structural representations. Then, we learn a refined representation across frames on the node-level via a dense graph kernel. The structural and temporal consistency of the graph forms the self-supervision of model training. HiGraph+ achieves excellent performance and demonstrates robustness in benchmark tests involving object, semantic part, keypoint, and instance labeling propagation tasks. Our algorithm implementations have been made publicly available at https://github.com/zyqin19/HiGraph.

Learning Semantic Keypoints for Object Detection in Aerial Images

Object detection in aerial images has achieved remarkable progress with the advent of deep convolutional neural networks (CNNs). It is, however, still a challenging task since the objects in aerial images are arbitrarily oriented and often densely packed. In this letter, we propose a novel method for oriented object detection in aerial images that represents objects as rotation equivariant semantic keypoints. Unlike conventional methods that represent object rotation according to angles from each axis in the Cartesian coordinate system, we represent object using a canonical orientation to ensure rotation equivariance. We accomplish this by representing an object as semantic keypoints, where each keypoint of the object consistently corresponds to the semantic part, regardless of rotation variation. To this end, we define the “head” point of the object as the canonical orientation and the remaining bounding box vectors as semantic keypoints in clockwise order. To discriminate visual attributes between different categories, we further use category-specific semantic keypoints, so that object classification and localization can be jointly solved in a cooperative manner. Our experiments demonstrate the effectiveness of rotation equivariant semantic keypoints on oriented object detection.

Структурно-композиционные приёмы инфотейнмента в медиатекстах региональных печатных СМИ

The novelty of the study is due to the need to systematize the techniques of infotainment used in the work of print media. The empirical research material for the work was the corpus of journalistic materials of the Belgorod regional youth magazine “OnOnas” for the period from 2013 to 2021. The relevance of the study lies in the need to theorize valuable empirical material and classify non-verbal methods of infotainment in relation to the practice of print media. The relevance and novelty of the study are also due to the insufficient study of the phenomenon of infotainment in relation to the practice of regional print media, since modern researchers still maintain a stable association of infotainment with television practices to the detriment of other types of media. The methods used in the course of the study include a system analysis with an emphasis on a constructive and structural-functional approach, and the typology method is also used. The conclusions are drawn that the use of non-verbal structural and compositional techniques of infotainment makes it possible to increase the visual appeal of journalistic material, include elements of entertainment for the purpose of informational unloading of the text, visualize statistical information or the semantic part of the text, captivate the audience with new forms of visualization and representation of journalistic information.

“Drink for the soul”: coffee tree and coffee as a symbol of the Arab national mentality

In the complex geopolitical situation of the beginning of the 21 century and with the change of the vector of interests towards the countries of the East, attention to all aspects of the life of the peoples of these countries is increasing due to the urgent need to understand their cultures, lifestyle, and the world around them. These peoples are increasingly active in the international arena, their role in various fields is becoming more noticeable, and it is important to interact with them at various levels. To do this, it is necessary to know what this people lives by, what their priority is, what is important or secondary for them. The complexity of intercultural communication lies in the fact that the national character and mentality of representatives of different linguistic cultures is based on the existing realities of a particular people’s life, its moral values and attitudes, through which various ethnic communities identify themselves and fix their place and purpose in this world. Symbols and images reflecting these ideas take either a material or verbal shell and convey an ethnic picture of the world of a particular people, which is completely different from the ideas of other ethnic groups about what surrounds them. To understand a representative of another linguistic consciousness, mastering the vocabulary and grammar of a foreign language may not be accurate enough. Significant elements of language in their semantic part also contain an extra-linguistic component, the socalled background knowledge, which can be gleaned from the history, mythology, literature, folklore of the people with whom we are interacting. This extra-linguistic component, according to the author, is of great importance for understanding the national mentality and national character. It helps to identify other driving forces of the actions of different peoples in the process of intercultural communication, along with their political and economic interests. In this article, the author considers one of the symbols of the national mentality of the Arabs – the coffee tree and its derivative – a coffee drink, which plays an important role in creating a national collective portrait of the Arabic society, as well as makes an attempt to show how and why they occupied such an important place in the Arab national consciousness.

INTERACTIVE STRATEGIES FOR ANALYZING G. BELGER’S NOVELS

This article discusses interactive strategies for analyzing G. Belger's novels. The writer's works are presented in the aspect of ideological and substantive strengthening of interethnic harmony and tolerance of the peoples of Kazakhstan. Understanding these aspects is the key to successful intercultural cooperation, the result of which is the peace and well-being of the peoples of our country. Kazakhstan is a Eurasian country, a bridge between diverse cultural, political and economic spaces. The work of Gerold Belger convincingly proves that for many decades a special semiosphere has been forming, including the language one; that each ethnic group has its own historical destiny and its own place in the ontology of humanity; that bilingualism, manifested in the creative process through translingualism, contributes to the organization of language systems more complex than "pure", contact-limited languages.
 The problem of translinguism - that is, the transfer by the author of the ethno-cultural content by means of the acquired language - is relatively new for the Kazakh nayka, but it is very relevant in the light of the trilingual policy chosen by our state for successful entry into the world com- munication. The translingual text suggests that the work of the author, ethnically representing one culture, was created in the language of another culture - acquired, adopted.
 In the transcultural analysis of the text, all levels of poetics are important - from architectonics to individual characters. Thus, the novel The House of the Wanderer Belger, written in Russian, has a structural organization that is somewhat atypical for a Russian novel. At the same time, it is not the three-part composition of the text itself that is unusual, but the fashionable breakdown of the novel, in which each of the semantic parts is the position of one of the characters and is named after him (David, Christian, Harry,). In this way, in our opinion, the principle of individualization of perception, which is typical for the German language picture of the world, is realized.
 In the literary life of Kazakhstan, Gerold Belger occupies a special place. Writer, critic, public figure, Belger creates in his work new images of the world, devoid of unification and cognitive simplicity. - that is, a schematic method of knowing being, which is based on a linear reaction to it.
 Trilingualism, which the writer metaphorically expressed by the formula three strings, became the resource base, thanks to which Herold Belger managed to create a model of the world that is more complex in heuristic and aesthetic terms for understanding these categories, we offer new methodological strategies for analyzing works of art.

Actualization of Subordinate Circumstantial Constructions of Adverbial Type in the Artistic Speech of Yevhen Gutsalo

The article characterizes the structural and semantic features of compound sentences with subordinate circumstantial parts of the adverbial type, which can be found in Yevhen Gutsal’s short prose. Both the semantic and syntactic relations between predicative parts, explicit connecting means are described. The functional load of the analyzed constructions in terms of artistic expression of the author’s approach is clarified. The productive use of compound sentences with adverbial circumstantial parts of spatial, temporal, methodical, quantitative and comparative semantics in the artistic speech of Yevhen Gutsalo is proved to be due to his communicative need to expand the syntactic level of time and space, quality and quantity. In general, compound sentences with proverbial subordinate clauses in the writer’s texts reflect the general tendencies in the construction and peculiarities of the functioning of these structures. The writer’s individual style of their use is manifested in the communicative need of these models to express specific semantic and syntactic relations, personal preferences for the choice of connecting means, the order of predicative parts, lexical and semantic content of structures. An important feature in the actualization of the analyzed sentences are the stylistic features of the writer’s language: lyricism, realism, colloquialism, psychologism, which requires attention to details, depiction of personal feelings and evaluation.

THE METHOD OF THEATRICALIZATION IN TEACHING YOUNGER STUDENTS TO SOLVE COMPOSITE TASKS

The method of theatricalization is described, which is used in solving textual mathematical problems in mathematics lessons in elementary grades at all stages of working with a composite problem. This improves students’ understanding of the structure of the task, its semantic parts, increases the likelihood that the task will be solved correctly, and there will be a versatile development of students.

Semantic speckle: an auto-located speckle pattern for DIC measurement.

Digital image correlation (DIC) has been widely used in both experimental mechanics and engineering fields. The matching algorithm of the DIC method usually requires surfaces containing a random speckle pattern as a deformation information carrier. The speckle pattern plays an irreplaceable role in DIC, which has led to extensive research on it. However, most previous research had always focused on the fabrication and computational performance of the speckle, ignoring the value of intentionally defining the meaning of speckle in design. In this study, we describe a novel, to the best of our knowledge, speckle pattern named semantic speckle. It is a digital speckle composed of several different speckle patterns with similar characteristics. Based on the deep-learning method and matching algorithm, the central location of the semantic part in the overall speckle image can be obtained automatically. Through the intentional definition of the semantic part, it can be possible to calibrate the camera parameters and correct the external parameters of the DIC systems.

PriFit: Learning to Fit Primitives Improves Few Shot Point Cloud Segmentation

AbstractWe present PriFit, a semi‐supervised approach for label‐efficient learning of 3D point cloud segmentation networks. PriFit combines geometric primitive fitting with point‐based representation learning. Its key idea is to learn point representations whose clustering reveals shape regions that can be approximated well by basic geometric primitives, such as cuboids and ellipsoids. The learned point representations can then be re‐used in existing network architectures for 3D point cloud segmentation, and improves their performance in the few‐shot setting. According to our experiments on the widely used ShapeNet and PartNet benchmarks, PriFit outperforms several state‐of‐the‐art methods in this setting, suggesting that decomposability into primitives is a useful prior for learning representations predictive of semantic parts. We present a number of ablative experiments varying the choice of geometric primitives and downstream tasks to demonstrate the effectiveness of the method.

A Masked Self-Supervised Pretraining Method for Face Parsing

Face Parsing aims to partition the face into different semantic parts, which can be applied into many downstream tasks, e.g., face mask up, face swapping, and face animation. With the popularity of cameras, it is easier to acquire facial images. However, pixel-wise manually labeling is time-consuming and labor-intensive, which motivates us to explore the unlabeled data. In this paper, we present a self-supervised learning method attempting to make full use of the unlabeled facial images for face parsing. In particular, we randomly mask some patches in the central area of facial images, and the model is required to reconstruct the masked patches. This self-supervised pretraining is capable of making the model capture facial feature representations through these unlabeled data. After self-supervised pretraining, the model is fine-tuned on a few labeled data for the face parsing task. Experimental results show that the model achieves better performance for face parsing assisted by the self-supervised pretraining, which greatly decreases the labeling cost. Our approach achieves 74.41 mIoU on the LaPa test set fine-tuned on only 0.2% of the labeled data of the whole training data, surpassing the model that is directly trained by a large margin of +5.02 mIoU. In addition, our approach achieves a new state-of-the-art on the LaPa and CelebAMask-HQ test set.

Part Decomposition and Refinement Network for Human Parsing

Dear Editor, This letter is concerned with human parsing based on part-wise semantic prediction. Human body can be regarded as a whole structure composed of different semantic parts, and the mainstream single human parser uses semantic segmentation pipeline to solve this problem. However, the differences between human parsing and semantic segmentation tasks bring some issues that are inevitable to avoid. In this paper, we propose a novel method called part decomposition and refinement network (PDRNet), which adopt part-wise mask prediction other than pixel-wise semantic prediction to tackle human parsing task. Specifically, we decompose the human body into different semantic parts and design a decomposition module to learn the central position of each part. The refinement module is proposed to obtain the mask of each human part by learning convolution kernel and convolved feature. In inference stage, the predicted human part masks are combined into a complete human parsing result. Through the decomposition, refinement and combination of human parts, PDRNet greatly reduces the confusion between the target human and the background human, and also significantly improves the semantic consistency of human part. Extensive experiments show that PDRNet performs favorably against state-of-the-art methods on several human parsing benchmarks, including LIP, CIHP and Pascal-Person-Part.

Which Conference Is That? A Case Study in Computer Science

Conferences play a major role in some disciplines such as computer science and are often used in research quality evaluation exercises. Differently from journals and books, for which ISSN and ISBN codes provide unambiguous keys, recognizing the conference series in which a paper was published is a rather complex endeavor: There is no unique code assigned to conferences, and the way their names are written may greatly vary across years and catalogs. In this article, we propose a technique for the entity resolution of conferences based on the analysis of different semantic parts of their names. We present the results of an investigation of our technique on a dataset of 42,395 distinct computer science conference names excerpted from the DBLP computer science repository, 1 which we automatically link to different authority files. With suitable data cleaning, the precision of our record linkage algorithm can be as high as 94%. A comparison with results obtainable using state-of-the-art general-purpose record linkage algorithms rounds off the article, showing that our ad hoc solution largely outperforms them in terms of the quality of the results.

Learning discriminative and generalizable features with multi-branch for person re-identification

Finer-grained local features play a supplementary role in the description of pedestrian global features, and the combination of them has been an essential solution to improve discriminative performances in person re-identification (PReID) tasks. The existing part-based methods mostly extract representational semantic parts according to human visual habits or some prior knowledge and focus on spatial partition strategies but ignore the significant influence of channel information on PReID task. So, we proposed an end-to-end multi-branch network architecture (MCSN) jointing multi-level global fusion features, channel features and spatial features in this paper to better learn more diverse and discriminative pedestrian features. It is worth noting that the effect of multi-level fusion features on the performance of the model is taken into account when extracting global features. In addition, to enhance the stability of model training and the generalization ability of the model, the BNNeck and the joint loss function strategy are applied to all vector representation branches. Extensive comparative evaluations are conducted on three mainstream image-based evaluation protocols, including Market-1501, DukeMTMC-ReID and MSMT17, to validate the advantages of our proposed model, which outperforms previous state-of-the-art in ReID tasks.

Multilabel learning based adaptive graph convolutional network for human parsing

In human parsing, graph convolutional networks (GCNs), which naturally model the skeleton of the human body as a fixed graph, have been witnessed to obtain remarkable performance. However, the existing methods perform the fixed graph modeling over all the training samples. This may not be an optimal graph for the diversity of the samples that contain various shapes of human parts, complex body postures, severe occlusions and dense crowd, etc. Focusing on this, we propose a new Multilabel Learning based Adaptive Graph Convolutional Network (ML-AGCN) for human parsing. The ML-AGCN includes three modules: adaptive graph generation module, semantic parts based attention module and label consistency loss. Concretely, to effectively deal with the different sizes and connectivities of the optimal graph for different samples, we first propose an adaptive graph generation module based on multilabel learning that contains graph node adaptation (GNA) and graph connection adaptation (GCA). Then, for a more comprehensive node embedding, we design a semantic parts based attention module to optimally fuse fixed graph embeddings and adaptive graph embeddings. Besides, to further explicitly constraint the consistency between the predicted multilabel and the predicted human parsing results, we propose a label consistency loss that can simultaneously refine the human parsing results and optimize the accuracy of the adaptive graph. Extensive experiments on four challenging datasets, including PASCAL-Person-Part, ATR, LIP and CIHP, well demonstrate the effectiveness of our model, and it outperforms other state-of-the-art methods in human parsing.

SofGAN: A Portrait Image Generator with Dynamic Styling

Recently, Generative Adversarial Networks (GANs) have been widely used for portrait image generation. However, in the latent space learned by GANs, different attributes, such as pose, shape, and texture style, are generally entangled, making the explicit control of specific attributes difficult. To address this issue, we propose a SofGAN image generator to decouple the latent space of portraits into two subspaces: a geometry space and a texture space. The latent codes sampled from the two subspaces are fed to two network branches separately, one to generate the 3D geometry of portraits with canonical pose, and the other to generate textures. The aligned 3D geometries also come with semantic part segmentation, encoded as a semantic occupancy field (SOF). The SOF allows the rendering of consistent 2D semantic segmentation maps at arbitrary views, which are then fused with the generated texturemaps and stylized to a portrait photo using our semantic instance-wise module. Through extensive experiments, we show that our system can generate high-quality portrait images with independently controllable geometry and texture attributes. The method also generalizes well in various applications, such as appearance-consistent facial animation and dynamic styling.

An Essay on Compositionality of Thoughts in Frege’s Philosophy

In the paper, I propose a novel approach to Frege’s view on the principle of compositionality, its relation to the propositional holism and the formation of concepts. The main idea is to distinguish three stages of constructing a logically perfect language. At the first stage, only a sentence as a whole expresses a Thought. It is impossible to assign meaning to less complex units. This is the stage of an ordinary language. The second phase concerns the proper level of construction of a logically perfect language. We are forced to discriminate syntactic and semantic parts of sentences to account for the inference relations. We can distinguish senses and references of parts of sentences. Furthermore, it is possible here to choose between different ways of analysing the given Thought. Finally, at the third stage, every expression of the language has an unambiguous sense and this sense determines a unique reference. The logically perfect language is ready. We may view Thoughts as composed from primitive elements. Moreover, the senses of parts of a sentence correspond to the parts of a Thought, so that the structure of the sentence serves as the image of the structure of the Thought. The principle of compositionality is met and we can discern how understanding of the infinite numbers of Thoughts is possible and how languages are learnable. The main advantage of the presented view is that it allows accommodating some aspects of Frege’s philosophy that are often seen as mutually incompatible. Furthermore, I submit extensive textual data in favour of the discussed views and conceptions.

SSAG-Net: Syntactic and Semantic Attention-Guided Machine Reading Comprehension

Machine reading comprehension (MRC) is a task in natural language comprehension. It assesses machine reading comprehension based on text reading and answering questions. Traditional attention methods typically focus on one of syntax or semantics, or integrate syntax and semantics through a manual method, leaving the model unable to fully utilize syntax and semantics for MRC tasks. In order to better understand syntactic and semantic information and improve machine reading comprehension, our study uses syntactic and semantic attention to conduct text modeling for tasks. Based on the BERT model of Transformer encoder, we separate a text into two branches: syntax part and semantics part. In syntactic component, an attention model with explicit syntactic constraints is linked with a self-attention model of context. In semantics component, after the framework semantic parsing, the lexical unit attention model is utilized to process the text in the semantic part. Finally, the vectors of the two branches converge into a new vector. And it can make answer predictions based on different types of data. Thus, a syntactic and semantic attention-guided machine reading comprehension (SSAG-Net) is formed. To test the model’s validity, we ran it through two MRC tasks on SQuAD 2.0 and MCTest, and the SSAG-Net model outperformed the baseline model in both.