Semantic Representation Research Articles

A method for named entity recognition in social media texts with syntactically enhanced multiscale feature fusion.

Social media data are characterized by significant noise and non-standardization, thereby posing challenges for existing methods in recognizing named entities owing to the entity sparsity and insufficient semantic richness. Thus, to deal with these issues, this study proposes SEMFF-NER, a named entity recognition (NER) method in social media texts that integrates multi-scale features and syntactic information. First, global features are extracted using a Transformer-based encoder (XLNET) with embedded dependency syntactic relations to enhance semantic representation. Next, sliding windows of different lengths capture local features, which are input into a bi-directional long short-term memory (BiLSTM) to capture multi-level local features. Subsequently, the fusion-attention mechanism effectively integrates global contextual information with multiple local features to predict the optimal entity labels. Extensive experiments conducted on three datasets collected from English social media platforms (WNUT2016, WNUT2017, OntoNotes5.0_English) demonstrate the advantageous performance of our proposed method, and ablation experiments further confirm the method's viability and effectiveness.

Unveiling user identity across social media: a novel unsupervised gradient semantic model for accurate and efficient user alignment

AbstractThe field of social network analysis has identified User Alignment (UA) as a crucial area of investigation. The objective of UA is to identify and connect user accounts across diverse social networks, even when there are no explicit interconnections. UA plays a pivotal role in synthesising coherent user profiles and delving into the intricacies of user behaviour across platforms. However, traditional approaches have encountered limitations. Singular embedding techniques have been found to fall short in fully capturing the semantic essence of user profile attributes. Furthermore, classification-based embedding methods lack definitive criteria for categorisation, thereby constraining both the efficacy and applicability of these models. This paper presents a novel unsupervised Gradient Semantic Model for User Alignment (GSMUA) for the purpose of identifying common user identities across social networks. GSMUA categorises user profile information into weak, sub, and strong gradients based on the semantic intensity of attributes. Different gradient semantic levels direct attention to literal features, semantic features, or a combination of both during feature extraction, thereby achieving a full semantic representation of user attributes. In the case of strongly semantic long texts, GSMUA employs Named Entity Recognition (ENR) technology in order to enhance the inefficient handling of such texts. Furthermore, GSMUA compensates for missing user profile attributes by utilising profile information from user neighbours, thereby reducing the negative impact of missing user profile attributes on model performance. Extensive experiments conducted on four pairs of real datasets demonstrate the superiority of our approach. In comparison to the most effective previously developed unsupervised methods, GSMUA demonstrates improvements in hit-precision ranging from 5.32 to 12.17%. When compared to supervised methods, the improvements range from 0.71 to 11.79%.

Right is it right? Influence of the type of motor response on behavioral and electrophysiological indicators during the orthographic decision task

The role of the sensorimotor component in the processing of verbal information is currently widely discussed. We hypothesize that the type of motor response may influence behavioral and electrophysiological performance in an orthographic decision task. Event-related potentials (ERPs) were recorded in 42 right-handers during an orthographic decision task. Half of the participants were instructed to press the right button with their right hand upon encountering correctly spelled words and the left button with their left hand upon encountering misspelled words, while the other half followed the opposite instruction. The motor response to correctly spelled words was shorter for the dominant hand compared to the non-dominant hand, which may be due to the coincidence of semantic and motor representations (stimulus-response compatibility effect). In addition, reaction times to incorrectly spelled words were longer than to correctly spelled words only in the group of participants who responded with their dominant hand to correctly spelled words. The P200 and N400 ERP components did not depend on the type of motor response. However, in the group of participants who pressed correctly spelled words with their right hand, the amplitude of the P600 component was greater for misspelled words compared to correctly spelled words. Thus, the type of motor response influenced the effects associated with word spelling recognition.

Activity in Occipito-Temporal Cortex Is Involved in Tool-Use Planning and Contributes to Tool-Related Semantic Neural Representations

Abstract Tool use and language are highly refined human abilities which may show neural commonalities due to their potential reciprocal interaction during evolution. Recent work provided evidence for shared neural resources between tool use and syntax. However, whether activity within the tool-use network also contributes to semantic neural representations of tool nouns remains untested. To this aim, we identified the tool-use planning network with functional magnetic resonance imaging while participants used pliers. The very same participants underwent a semantic priming task including two categories, tool nouns and animal nouns, to highlight the respective underlying networks. With multivariate analyses of the activation neural patterns, we tested whether activity in tool-use brain clusters takes part in the neural representation of tool nouns as compared with animal nouns. The results revealed that word semantic categories were decoded within the left occipito-temporal cortex activated by preparing to use a tool, with similar patterns of brain activity for words within the same category. In addition, in the same area, neural activations for tool nouns were found to be higher than those for animal nouns. These findings suggest that activity in tool-use related brain areas encodes semantic information separately for tool nouns and animal nouns, thus supporting the embodiment of tool-noun processing in the tool-use sensorimotor network.

AWA-GCN: Enhancing Chinese sentiment analysis with a novel GCN model for Triplet and Quadruplet Extraction at SIGHAN 2024 dimABSA Task

This study presents the AWA-GCN (Attention-Weighted Affine with Graph Convolutional Networks) model, a new method designed exclusively for the SIGHAN 2024 Chinese Dimensional Aspect-Based Sentiment Analysis (dimABSA) Task. This test requires participants to identify sentiment triples and sentiment quadruples from textual data. The data comprises aspects, sentiment views, categories, and intensities across valence-arousal dimensions. Our model differs from existing models by utilizing a complex multi-layer attention mechanism within a GCN architecture, instead of relying on pipeline techniques or Grid Tagging Schemes (GTS). This design successfully captures the intricate interconnections among many sentiment components. The AWA-GCN model is very innovative in its ability to effectively handle quadruple extraction, which is a complex problem that has not been handled by traditional methods. Our approach enhances the understanding of aspect-sentiment interactions by including sophisticated techniques like word-level attention and semantic graph representations. The AWA- GCN model outperforms previous baselines in precision, recall, and F1 score on the dimABSA dataset, as demonstrated by empirical assessments. Additionally, the model exhibits significant enhancements in capturing the dimensional features of sentiment expressions. The results validate the models exceptional ability to handle the complex nuances necessary for successful sentiment analysis in simple Chinese.

A Hybrid Semantic Representation Method Based on Fusion Conceptual Knowledge and Weighted Word Embeddings for English Texts

The accuracy of traditional topic models may be compromised due to the sparsity of co-occurring vocabulary in the corpus, whereas conventional word embedding models tend to excessively prioritize contextual semantic information and inadequately capture domain-specific features in the text. This paper proposes a hybrid semantic representation method that combines a topic model that integrates conceptual knowledge with a weighted word embedding model. Specifically, we construct a topic model incorporating the Probase concept knowledge base to perform topic clustering and obtain topic semantic representation. Additionally, we design a weighted word embedding model to enhance the contextual semantic information representation of the text. The feature-based information fusion model is employed to integrate the two textual representations and generate a hybrid semantic representation. The hybrid semantic representation model proposed in this study was evaluated based on various English composition test sets. The findings demonstrate that the model presented in this paper exhibits superior accuracy and practical value compared to existing text representation methods.

Cross-language facilitatory and inhibitory effects in the naming of Japanese words by Chinese-Japanese bilinguals

Abstract This study investigates the effects of orthographic and phonological overlaps between the first language (L1) and the second language (L2) during word naming by adult Chinese-Japanese bilinguals, in the light of the Orthographic Depth Hypothesis which posits that the processing of printed words depends on the extent of grapheme-phoneme correspondence in the language. We compared their naming performance of L2 Japanese words written in moraic hiragana which is not used in their L1 Chinese, morphosyllabic kanji utilized in L1, and a mixed condition. Results indicated a longer latency for naming L2 Japanese words with kun-reading which has no phonological overlap with L1, particularly when written in kanji, which has orthographic overlap with L1. Consistent with the Orthographic Depth Hypothesis, participants performed better for words in hiragana as a shallow orthography than for words in kanji as a deep orthography. Furthermore, this study examined the priming effect of L2 Japanese translation equivalents on L1 Chinese targets. We found more accurate and faster naming of L1 words semantically exposed by L2, suggesting that L2 semantic representation facilitates L1 translation equivalence.

MAMSC: a semantic enhanced representation model for public opinion key node recognition based on multianchor mapping in semantic communities

PurposeInformation is presented in various modalities such as text and images, and it can quickly and widely spread on social networks and among the general public through key communication nodes involved in public opinion events. Therefore, by tracking and identifying key nodes of public opinion, we can determine the direction of public opinion evolution and timely and effectively control public opinion events or curb the spread of false information.Design/methodology/approachThis paper introduces a novel multimodal semantic enhanced representation based on multianchor mapping semantic community (MAMSC) for identifying key nodes in public opinion. MAMSC consists of four core components: multimodal data feature extraction module, feature vector dimensionality reduction module, semantic enhanced representation module and semantic community (SC) recognition module. On this basis, we combine the method of community discovery in complex networks to analyze the aggregation characteristics of different semantic anchors and construct a three-layer network module for public opinion node recognition in the SC with strong, medium and weak associations.FindingsThe experimental results show that compared with its variants and the baseline models, the MAMSC model has better recognition accuracy. This study also provides more systematic, forward-looking and scientific decision-making support for controlling public opinion and curbing the spread of false information.Originality/valueWe creatively combine the construction of variant autoencoder with multianchor mapping to enhance semantic representation and construct a three-layer network module for public opinion node recognition in the SC with strong, medium and weak associations. On this basis, our constructed MAMSC model achieved the best results compared to the baseline models and ablation evaluation models, with a precision of 91.21%.

Morpho-orthographic segmentation on visual word recognition in Brazilian Portuguese speakers

ObjectiveStudies have demonstrated a morphological decomposition process in the initial stages of visual word recognition based on orthographically defined morphemes and independent of semantic and lexical information. The present study sought to investigate this phenomenon in adult Brazilian Portuguese-speaking readers.MethodParticipants performed a lexical decision task preceded by primes at two Stimulus Onset Asynchrony (SOA): 33 or 250 ms. The primes had one of the following relations with the target word: morphological (porteiro [DOORKEEPER]–PORTA [DOOR]), morpho-orthographic (cordeiro [LAMB]–CORDA [ROPE]), orthographic (abril [APRIL]–ABRIU [OPENED]), semantic (abelha [BEE]–MEL [HONEY]), or no relation (pessoa [PERSON]–DADO [DICE]).ResultsPriming effects were observed for the morphological and semantic conditions at both SOAs but not for the orthographic and morpho-orthographic conditions.ConclusionOur results suggest that semantic representations mediate morphological priming in Brazilian Portuguese since the early stages of visual word recognition.

Discovery of optimal cell type classification marker genes from single cell RNA sequencing data

BackgroundThe use of single cell/nucleus RNA sequencing (scRNA-seq) technologies that quantitively describe cell transcriptional phenotypes is revolutionizing our understanding of cell biology, leading to new insights in cell type identification, disease mechanisms, and drug development. The tremendous growth in scRNA-seq data has posed new challenges in efficiently characterizing data-driven cell types and identifying quantifiable marker genes for cell type classification. The use of machine learning and explainable artificial intelligence has emerged as an effective approach to study large-scale scRNA-seq data.MethodsNS-Forest is a random forest machine learning-based algorithm that aims to provide a scalable data-driven solution to identify minimum combinations of necessary and sufficient marker genes that capture cell type identity with maximum classification accuracy. Here, we describe the latest version, NS-Forest version 4.0 and its companion Python package (https://github.com/JCVenterInstitute/NSForest), with several enhancements to select marker gene combinations that exhibit highly selective expression patterns among closely related cell types and more efficiently perform marker gene selection for large-scale scRNA-seq data atlases with millions of cells.ResultsBy modularizing the final decision tree step, NS-Forest v4.0 can be used to compare the performance of user-defined marker genes with the NS-Forest computationally-derived marker genes based on the decision tree classifiers. To quantify how well the identified markers exhibit the desired pattern of being exclusively expressed at high levels within their target cell types, we introduce the On-Target Fraction metric that ranges from 0 to 1, with a metric of 1 assigned to markers that are only expressed within their target cell types and not in cells of any other cell types. NS-Forest v4.0 outperforms previous versions in simulation studies and on its ability to identify markers with higher On-Target Fraction values for closely related cell types in real data, and outperforms other marker gene selection approaches for cell type classification with significantly higher F-beta scores when applied to datasets from three human organs—brain, kidney, and lung.DiscussionFinally, we discuss potential use cases of the NS-Forest marker genes, including for designing spatial transcriptomics gene panels and semantic representation of cell types in biomedical ontologies, for the broad user community.

Semantic layout-guided diffusion model for high-fidelity image synthesis in ‘The Thousand Li of Rivers and Mountains’

“The Thousand Li of Rivers and Mountains” (TLRM) is one of the most famous traditional paintings. Even experienced artists face significant challenges in replicating TLRM. Utilizing semantic layout-guided diffusion models offers a novel approach to this task; however, synthesizing high-quality images with consistent semantics and layout is difficult due to TLRM’s unique characteristics. In this paper, we propose a novel diffusion model-based framework guided by semantic layout to generate images tailored to TLRM’s distinct features. We introduce the layout enhanced map and latent layout injection strategies, which improve semantic fidelity and color distribution. These innovations are integrated into our semantic latent diffusion model for effective semantically-guided image generation. To address the training challenges posed by a single large-scale image, we employ specialized data augmentation techniques, facilitating the generation of one-hot semantic layout representations and ensuring continuous training until convergence. Additionally, we developed the TLRM-dataset for semantic image synthesis, which enhances visual quality and semantic consistency. Experimental results and user surveys demonstrate that our model can produce high-fidelity and diversified results, offering significant potential for art education, digital preservation, and cultural heritage. Our code is publicly available at https://github.com/rane7/TLRM-SLDM.

Superpixel semantics representation and pre-training for vision-language tasks

Superpixel semantics representation and pre-training for vision-language tasks

Boosting multi-document summarization with hierarchical graph convolutional networks

The input of the multi-document summarization task is usually long and has high redundancy. Encoding multiple documents is a challenge for the Seq2Seq architecture. The way of concatenating multiple documents into a sequence ignores the relation between documents. Attention-based Seq2Seq architectures have slightly improved the cross-document relation modeling for multi-document summarization. However, these methods ignore the relation between sentences, and there is little improvement that can be achieved through the attention mechanism alone. This paper proposes a hierarchical approach to leveraging the relation between words, sentences, and documents for abstractive multi-document summarization. Our model employs the Graph Convolutional Networks (GCN) for capturing the cross-document and cross-sentence relations. The GCN module can enrich semantic representations by generating high-level hidden features. Our model achieves significant improvement over the attention-based baseline, beating the Hierarchical Transformer by 3.4/1.64, 1.92/1.44 ROUGE-1/2 F1 points on the Multi-News and WikiSum datasets, respectively. Experimental results demonstrate that our delivered method brings substantial improvements over several strong baselines.

Between compounding and phrasal derivation: Polish complex nouns in sam(o)-

The aim of this paper is to explore the class of Polish complex nouns with the initial combining form samo-, such as samobójca ‘suicide’. The study seeks to determine the morphological structure of what we consider to be a case of ‘syntax-inside-morphology’ in the domain of Polish word-formation. We want to argue that what appears to be sheer concatenation of morphological building blocks, often routinely classified as a compound word with a lefthand bound root, could, in fact, be interpreted as a derived noun based on a phrasal unit. In what follows, we shall argue that the morphosemantics of Polish complex nouns in samo- can successfully be captured by left-branching structure, which ensures that their morphological and semantic representations cohere into a well-organised whole. The dispute over derivational phenomena exhibiting a degree of reciprocity between morphology and syntax has long surrounded the data-driven examination of certain word-formation types across languages. Currently, a rich body of linguistic data amassed across various languages seems to strongly undermine the aprioristic separation of the two modules of grammar. The present study draws on tenets central to theories which are neither strictly lexicalist nor syntactocentric, in particular the Firewall Theory advanced by Lieber & Scalise (2006) .

Joint embedding in Hierarchical distance and semantic representation learning for link prediction

The link prediction task aims to predict missing entities or relations in the knowledge graph and is essential for the downstream application. Existing well-known models deal with this task by mainly focusing on representing knowledge graph triplets in the distance space or semantic space. However, they can not fully capture the information of head and tail entities, nor even make good use of hierarchical level information. Thus, in this paper, we propose a novel knowledge graph embedding model for the link prediction task, namely, HIE, which models each triplet (h, r, t) into distance measurement space and semantic measurement space, simultaneously. Moreover, HIE is introduced into hierarchical-aware space to leverage rich hierarchical information of entities and relations for better representation learning. Specifically, we apply distance transformation operation on the head entity in distance space to obtain the tail entity instead of translation-based or rotation-based approaches. Experimental results of HIE on four real-world datasets show that HIE outperforms several existing state-of-the-art knowledge graph embedding methods on the link prediction task and deals with complex relations accurately.

Addressing Semantic Variability in Clinical Outcome Reporting Using Large Language Models

Background/Objectives: Clinical trials frequently employ diverse terminologies and definitions to describe similar outcomes, leading to ambiguity and inconsistency in data interpretation. Addressing the variability in clinical outcome reports and integrating semantically similar outcomes is important in healthcare and clinical research. Variability in outcome reporting not only hinders the comparability of clinical trial results but also poses significant challenges in evidence synthesis, meta-analysis, and evidence-based decision-making. Methods: This study investigates variability reduction in outcome measures reporting using rule-based and large language-based models. It aims to mitigate the challenges associated with variability in outcome reporting by comparing these two models. The first approach, which is rule-based, will leverage well-known ontologies, and the second approach exploits sentence-bidirectional encoder representations from transformers (SBERT) to identify semantically similar outcomes along with Generative Pre-training Transformer (GPT) to refine the results. Results: The results show that the relatively low percentages of outcomes are linked to established rule-based ontologies. Analysis of outcomes by word count highlighted the absence of ontological linkage for three-word outcomes, which indicates potential gaps in semantic representation. Conclusions: Employing large language models (LLMs), this study demonstrates its ability to identify similar outcomes, even with more than three words, suggesting a crucial role in outcome harmonization efforts, potentially reducing redundancy and enhancing data interoperability.

Two languages, two emotional minds in one brain: processing emotion-label and emotion-laden words by Chinese-English bilinguals

ABSTRACT The present study examined the neural correlates of emotion effects evoked by emotion-label and emotion-laden nouns in Chinese-English bilinguals’ two languages through the emotion categorization tasks. At the perceptual processing stage, only L2 emotion-label and emotion-laden nouns induced amplified N100 than neutral nouns. At the semantic processing stage, relative to neutral nouns, L2 negative emotion-laden nouns triggered heightened EEPN, and L1 emotion-laden nouns evoked marginally smaller EEPN; only L2 emotion-laden and positive emotion-label nouns exhibited augmented LEPN; L2 positive emotion-label and L1 emotion-label and emotion-laden nouns displayed reduced N400. At the integration processing stage, L2 positive nouns showed larger LPC and L1 positive nouns displayed smaller LPC than neutral nouns. These results verified emotion word type hypothesis across languages, and also implied that emotional attributes were encoded within both perceptual and semantic representations in L2, and predominantly within semantic representation in L1.

JobFormer: Skill-Aware Job Recommendation with Semantic-Enhanced Transformer

Job recommendation aims to provide potential talents with suitable job descriptions (JDs) consistent with their career trajectory, which plays an essential role in proactive talent recruitment. In real-world management scenarios, the available JD-user records always consist of JDs, user profiles, and click data, in which the user profiles are typically summarized as the user's skill distribution for privacy reasons. Although existing sophisticated recommendation methods can be directly employed, effective recommendation still has challenges considering the information deficit of JD itself and the natural heterogeneous gap between JD and user profile. To address these challenges, we proposed a novel skill-aware recommendation model based on the designed semantic-enhanced Transformer to parse JDs and complete personalized job recommendation. Specifically, we first model the relative items of each JD and then adopt an encoder with the local-global attention mechanism to better mine the intra-job and inter-job dependencies from JD tuples. Moreover, we adopt a two-stage learning strategy for skill-aware recommendation, in which we utilize the skill distribution to guide JD representation learning in the recall stage, and then combine the user profiles for final prediction in the ranking stage. Consequently, we can embed rich contextual semantic representations for learning JDs, while skill-aware recommendation provides effective JD-user joint representation for click-through rate (CTR) prediction. To validate the superior performance of our method for job recommendation, we present a thorough empirical analysis of large-scale real-world and public datasets to demonstrate its effectiveness and interpretability.

MMLmiRLocNet: miRNA Subcellular Localization Prediction based on Multi-view Multi-label Learning for Drug Design.

Identifying subcellular localization of microRNAs (miRNAs) is essential for comprehensive understanding of cellular function and has significant implications for drug design. In the past, several computational methods for miRNA subcellular localization is being used for uncovering multiple facets of RNA function to facilitate the biological applications. Unfortunately, most existing classification methods rely on a single sequencebased view, making the effective fusion of data from multiple heterogeneous networks a primary challenge. Inspired by multi-view multi-label learning strategy, we propose a computational method, named MMLmiRLocNet, for predicting the subcellular localizations of miRNAs. The MMLmiRLocNet predictor extracts multi-perspective sequence representations by analyzing lexical, syntactic, and semantic aspects of biological sequences. Specifically, it integrates lexical attributes derived from k-mer physicochemical profiles, syntactic characteristics obtained via word2vec embeddings, and semantic representations generated by pre-trained feature embeddings. Finally, module for extracting multi-view consensus-level features and specific-level features was constructed to capture consensus and specific features from various perspectives. The full connection networks are utilized as the output module to predict the miRNA subcellular localization. Experimental results suggest that MMLmiRLocNet outperforms existing methods in terms of F1, subACC, and Accuracy, and achieves best performance with the help of multi-view consensus features and specific features extract network.

Distributional hypothesis as isomorphism between word-word co-occurrence and analogical parallelograms.

Most of the modern natural language processing (NLP) techniques are based on the vector space models of language, in which each word is represented by a vector in a high dimensional space. One of the earliest successes was demonstrated by the four-term analogical reasoning task: what is to C as B is to A? The trained word vectors form "parallelograms" representing the quadruple of words in analogy. This discovery in NLP offers us insight into our understanding of human semantic representation of words via analogical reasoning. Despite successful applications of the large-scale language models, it has not been fully understood why such parallelograms emerge by learning through natural language data. As the vector space model is not optimized to form parallelograms, the key structure related to geometric shapes of word vectors is expected to be in the data, rather than the models. In the present article, we test our hypothesis that such parallelogram arrangement of word vectors readily exists in the co-occurrence statistics of language. Our approach focuses more on the data itself, and it is different from the existing theoretical approach trying to find the mechanism of parallelogram formation in the algorithms and/or vector arithmetic operations on word vectors. First, our analysis suggested that analogical reasoning is possible by decomposition of the bigram co-occurrence matrix. Second, we demonstrated the formation of a parallelepiped, a more structured geometric object than a parallelogram, by creating a small artificial corpus and its word vectors. With these results, we propose a refined form of distributional hypothesis pointing out an isomorphism between a sort of symmetry or exchangeability and word co-occurrence statistics.