Graph Neural Collaborative Topic Model for Citation Recommendation

Due to the overload of published scientific articles, citation recommendation has long been a critical research problem for automatically recommending the most relevant citations of given articles. Relational topic models (RTMs) have shown promise on citation prediction via joint modeling of document contents and citations. However, existing RTMs can only capture pairwise or direct (first-order) citation relationships among documents. The indirect (high-order) citation links have been explored in graph neural network–based methods, but these methods suffer from the well-known explainability problem. In this article, we propose a model called Graph Neural Collaborative Topic Model that takes advantage of both relational topic models and graph neural networks to capture high-order citation relationships and to have higher explainability due to the latent topic semantic structure. Experiments on three real-world citation datasets show that our model outperforms several competitive baseline methods on citation recommendation. In addition, we show that our approach can learn better topics than the existing approaches. The recommendation results can be well explained by the underlying topics.

INTERJECTIVE CONSTRUCTIONS WITH COMPONENT ADA IN THE TURKIC AND MONGOLIAN LANGUAGES: ON THE NATURE OF INTERJECTIONS

Статья посвящена описанию феномена тюркско - монгольской лексемы ада, пока ещё спорной в плане исторической принадлежности, обладающей гибкостью и сложностью семантической структуры. В монгольских и якутском языках ад (также адьарай/ажарай) является полноценной единицей с обобщённой семантикой «демон, злой дух». В тюркских языках слово ада действует как компонент междометных конструкций, имеющих семантические, грамматические и синтаксические особенности. Ада в них полностью десемантизировался, сохранив свои грамматические свойства. Конструкции с ада выражают различные эмоции: удивление, восторг, восхищение и т. д. Лингвоспецифичность хакасских междометных конструкций заключается в том, что они представляют собой реакцию говорящего на чьё - либо высказывание, оцененное как абсурдное, недостоверное. The article is devoted to the description of the phenomenon of the Turkic-Mongolian lexeme of hell, which is still controversial in terms of historical affiliation, and has a flexible and complex semantic structure. In the Mongolian and Yakut languages, ad (also adyaray / azharay) is a full - fledged unit with the generalized semantics of 'demon, evil spirit'. In the Turkic languages, word ada acts as a component of interjective constructions that have semantic, grammatical and syntactic features. Ada in them is completely desemanticized, retaining its grammatical properties. Constructions with ada express various emotions: surprise, delight, admiration, etc. The linguistic specificity of Khakass interjective constructions is that they represent a speaker's reaction to someone's statement, evaluated as absurd, unreliable.

Throwaway Shadows Using Parallel Encoders Generative Adversarial Network

Face photographs taken on a bright sunny day or in floodlight contain unnecessary shadows of objects on the face. Most previous works deal with removing shadow from scene images and struggle with doing so for facial images. Faces have a complex semantic structure, due to which shadow removal is challenging. The aim of this research is to remove the shadow of an object in facial images. We propose a novel generative adversarial network (GAN) based image-to-image translation approach for shadow removal in face images. The first stage of our model automatically produces a binary segmentation mask for the shadow region. Then, the second stage, which is a GAN-based network, removes the object shadow and synthesizes the effected region. The generator network of our GAN has two parallel encoders—one is standard convolution path and the other is a partial convolution. We find that this combination in the generator results not only in learning an incorporated semantic structure but also in disentangling visual discrepancies problems under the shadow area. In addition to GAN loss, we exploit low level L1, structural level SSIM and perceptual loss from a pre-trained loss network for better texture and perceptual quality, respectively. Since there is no paired dataset for the shadow removal problem, we created a synthetic shadow dataset for training our network in a supervised manner. The proposed approach effectively removes shadows from real and synthetic test samples, while retaining complex facial semantics. Experimental evaluations consistently show the advantages of the proposed method over several representative state-of-the-art approaches.

Action Verb Nosi `Cooking` in Bima Language: A Study of Natural Semantic Metalanguage

There are two subordinate structures in action verbs, namely uncompositional polysemy. Action verbs to do and happen are action verbs with uncomposed polysemy, including Cooking Nose. This research aims to find out the mapping and explication of the 'cooking' action verb. The method used is qualitative. The data source consists of oral data taken from key informants, written data collected from storybooks in the Bima language and language intuition. The technique used is interview and literature study. The data collection method is advanced, namely the agih method with the application of transformation and insertion techniques used to reveal the original meaning contained in BBM. The default meaning is used to determine the semantic structure of VBBm by explication or paraphrasing techniques. Each verb nosi cooking is based on: the tool, the model of movement, the part of the entity that is being treated, the result that the agent wants to achieve. The results show that the Bima language action verb 'Cooking' in general has a component mapping `X Doing something to Y` and therefore `Something happened to Y`. A number of words that contain the meaning of cooking: lowi, mbako, danda, salunga, puru, sanggowo, sanggapi, suje, ncango, and tumi, gule. This variant has unique semantic characteristics so that the meaning content of each word is different even though it is still in the same field of meaning (Cf. Adawiyah, 2021).

Biblical Etymology of Relatives: Origin of Language

Regarding origin of language, Genesis claims that ancient languages were divinely diversified. This testimony presents systematic evidence for biblical etymology related to relatives, solving word riddles on all relatives. Family is flame of altar, and relative is related to altar, because early relatives were worshipers around altar. Patriotism is parental worship, patriot is to cut in pair at altar, aunt is to count, offspring is off spring, descendant is related to census, blessing is related to sibling as brother, sister is to assist, woman is own, female is self in flesh, to marry is to Mary, to marry is to arrive, bride is related to tribe, Child is under cloud in addition to Infant’s finance, nephew has new place or no female, and girl and ugly are not to offer with hand in unclean period. 息 breath has rest息. 休 rest has tree木. Testimony has stone of ten commandments in tent . This major discovery in civilization on multilingual mutual match to biblical etymology addresses fundamental semantic structure of words in relatives. This provides revolutionary method of learning vocabulary by bilingual mutual match to biblical etymology

Oceny w opisie znaczenia jednostek leksykalnych – wybrane problemy

Autorka artykułu stawia tezę, że po pierwsze, czynności związane z oceną kogoś lub czegoś są prymarnie lingwistyczne. Po drugie, ocena w ściśle aksjologicznym sensie tego słowa powinna być odróżniona od wartościowań i innych oszacowań. Po trzecie, oceny zawartej w znaczeniu jednostki leksykalnej nie należy zrównywać z rozmaitymi cechami, które ta jednostka konotuje pragmatycznie. Po czwarte, błędem jest identyfikowanie komponentów oceniających z efektami emotywnymi lub ekspresywnymi. Cechy semantyczne związane z ocenami mogą być zlokalizowane w różnych miejscach struktury semantycznej: zarówno w jej tematycznej, jak i rematycznej części. W słownikach wygodnie jest opisywać oceny dane rematycznie przy użyciu stosownych kwalifikatorów, np. pejor ‘pejoratywne’ i amelior ‘amelioratywne’. Evaluations in the semantic description of lexical units: selected problems Summary: The present author puts forward the thesis that (1) activities related to the evaluation of someone or something are primarily linguistic, (2) an evaluation in the strictly axiological sense of the word should be distinguished from other estimates, (3) an evaluation embedded in the meaning of a lexical unit should not be equated with the various features that this unit connotes pragmatically and (4) it is wrong to identify evaluation components with emotive or expressive effects. The semantic features related to evaluations can be located in various places within the semantic structure of an expression, i.e. in both its thematic and rhematic parts. In dictionaries, it is convenient to employ rhematic evaluations which make use of appropriate qualifiers, e.g. pejor ‘pejorative’ and amelior ‘ameliorative’.