Semantic Noise Research Articles

Decoupled semantic graph neural network for knowledge graph embedding

Knowledge graph embedding (KGE) learns an embedding space for a more accurate representation of entities and relations. Although the KGE model has proliferated, it often fails to fully capture the rich semantics in knowledge graphs. Some studies attempt to apply decoupling methods to decompose node representations, yet they neglect the semantic noise generated during the decoupled process. To address these issues, we introduce the decoupled semantic graph network for KGE (named DSGNet), which employs a novel approach that combines semantic decoupling with structural awareness. DSGNet begins by projecting the knowledge graph into distinct semantic spaces, ensuring minimal correlation between them to achieve effective semantic decoupling. To mitigate semantic noise, a top-k sampling technique is applied to the decoupled graphs. DSGNet then aggregates these different semantic graphs using a relation-aware aggregation module, followed by a multi-layer aggregation process for enhanced node representation. The extensive experiments demonstrate that DSGNet performs competitively on two popular datasets. We make our code publicly available at https://github.com/HubuKG/DSGNet.

Development of the noise questionnaire in the online learning process and implications for counseling

Starting with online learning requires a lot of attention such as focus, comfortable sitting, and avoiding distractions such as noise. A noise questionnaire instrument is an evaluation tool designed to identify the experience of students in the online learning process. This instrument was developed based on a literature review on noise and online learning. The instrument first stage was given to 110 students and the instrument second stage was given to 460 students in seven universities in Indonesia, 99 male and 361 female respondents aged 18-30 years. The instrument was designed based on DeVito’s noise theory: physical noise, physiological noise, psychological noise, and semantic noise. The statistical test of the instrument used confirmatory factor analysis (CFA) to find the goodness of fit index model. The results of the noise instrument factor analysis show a fit model, acceptable validity, and high internal consistency (α=0.86). The findings of this study produce valid and reliable instruments for identifying noise indicators that are dominant in online learning activities. The results of identifying noise in online learning can be used to design guidance and counseling programs or plan actions to deal with noise in online learning according to the data obtained.

KOMUNIKASI INTERPERSONAL GURU DAN MURID DALAM MENGIMPLEMENTASIKAN TRI KAYA PARISUDHA DI SANGGAR KEGIATAN BELAJAR KOTA DENPASAR

In the implementation of character education in educational institutions, interpersonal communication patterns are very important so that character education can run well. This is because each student has a different background. This research was conducted to examine the interpersonal communication of teachers and students in implementing tri kaya parisudha at the Learning Activity Center in Denpasar. The theory that underlies this research is Rational Choice Theory. This study used a qualitative approach. The type of qualitative data collected is sourced from primary and secondary data. Data collection techniques used are observation, interviews, documentation. The findings of the research results are that the interpersonal communication applied by the teacher to the students of the Denpasar City Learning Activity Center in forming the religious character of children is using the preservation strategy model, one-way communication and interaction communication or two-way communication. This type of communication is expected to foster student activity both in teaching and learning activities and in play activities. The obstacles encountered are mechanical constraints (mechanical, channel noise), semantic between students and teachers (semantic noise), ecological constraints in learning activities in the city of Denpasar. Quality improvement is carried out by increasing educational input, improving the quality of the educational process, improving the school environment. The implications of interpersonal communication between teachers and students in tri kaya parisudha learning in the Learning Activity Center in Denpasar city are changes in the quality of children from the cognitive aspect, changes in the quality of children from the affective aspect, changes in the quality of children from the psychomotor aspect.

A Community Detection Method to Counter the Semantic Noise of Complex Networks: Bridging a Topology and Semantic Subspace Transformation

A Community Detection Method to Counter the Semantic Noise of Complex Networks: Bridging a Topology and Semantic Subspace Transformation

Autoencoder based image quality metric for modelling semantic noise in semantic communications

AbstractSemantic communication has attracted significant attention as a key technology for emerging 6G communications. This paper proposes an autoencoder based image quality metric to quantify the semantic noise. An autoencoder is initially trained with the reference image to generate the encoder‐decoder model and calculate its Latent Vector Space (LVS) and then a semantically generated/received image is inserted into the same autoencoder to create the corresponding LVS. Finally, both LVS are used to define the Euclidean space to calculate the mean square error between two LVS. Results indicate that the proposed model has a high correlation coefficient of 88% with subjective quality assessment and commonly used conventional metrics completely failed in semantic noise modelling.

Variation-aware semantic image synthesis

Semantic image synthesis (SIS) aims to produce photorealistic images aligning to given conditional semantic layout and has witnessed a significant improvement in recent years. Although the diversity in image-level has been discussed heavily, class-level mode collapse widely exists in current algorithms. Therefore, we declare a new requirement for SIS to achieve more photorealistic images, variation-aware, which consists of inter- and intra-class variation. The inter-class variation is the diversity between different semantic classes while the intra-class variation stresses the diversity inside one class. Through analysis, we find that current algorithms elusively embrace the inter-class variation but the intra-class variation is still not enough. Further, we introduce two simple methods to achieve variation-aware semantic image synthesis (VASIS) with a higher intra-class variation, semantic noise and position code. We combine our method with several state-of-the-art algorithms and the experimental result shows that our models generate more natural images and achieve slightly better FIDs and/or mIoUs than the counterparts. Our codes and models will be publicly available.

Robust Semantic Communications With Masked VQ-VAE Enabled Codebook

Although semantic communications have exhibited satisfactory performance on a large number of tasks, the impact of semantic noise and the robustness of the systems have not been well investigated. Semantic noise refers to the misleading between the intended semantic symbols and received ones, thus causes the failure of tasks. In this paper, we first propose a framework for the robust end-to-end semantic communication systems to combat the semantic noise. In particular, we analyze sample-dependent and sample-independent semantic noise. To combat the semantic noise, the adversarial training with weight perturbation is developed to incorporate the samples with semantic noise in the training dataset. Then, we propose to mask a portion of the input, where the semantic noise appears frequently, and design the masked vector quantized-variational autoencoder (VQ-VAE) with the noise-related masking strategy. We use a discrete codebook shared by the transmitter and the receiver for encoded feature representation. To further improve the system robustness, we develop a feature importance module (FIM) to suppress the noise-related and task-unrelated features. Thus, the transmitter simply needs to transmit the indices of these important task-related features in the codebook. Simulation results show that the proposed method can be applied in many downstream tasks and significantly improve the robustness against semantic noise with remarkable reduction on the transmission overhead.

Older people with mild cognitive impairment exhibit lower semantic noise after six months of frequent social conversations

AbstractBackgroundSemantic noise (SN) is a linguistic measurement that quantifies how well the conversation focuses on topics they are discussing. We hypothesized that by engaging in high frequent conversation which aimed to enhance cognitive reserve, the SN observed in older people’s conversations with mild cognitive impairment (MCI) would resemble those with normal cognition (NC).MethodWe analyzed transcriptions from a behavioral intervention randomized controlled trial, I‐CONECT (NCT02871921), which involved semi‐structured conversations between facilitators and participants. In the project, the experimental group engaged in 30‐minute semi‐structured conversations (four times per week for 6 months) via video‐chats. Every video‐chat is oriented to a predefined topic which was provided in a standardized order (e.g., dancing is the 1st conversation topic in the 2nd week). Out of 186 subjects randomized into the trial, 64 participants were randomized into the experimental group. Among them, 52 participants (26 MCI and 26 normal based on NACC D1) with sufficient recording quality during the 2nd week of the intervention (baseline) and during the last week (i.e., 24th) before the month 6 (M6) post‐intervention assessments (post‐trial assessments at M6) were used in this analysis. We utilized an automatic speech recognition (ASR) system to transcribe these recordings and measured each transcription’s SN. We analyzed the statistical difference of MSN (mean SN) between participants with MCI and those with NC at baseline (2nd week) and M6 (24th week).ResultsUsing two‐tailed t‐test, we found that at the baseline, MSN in MCI group was significantly higher than NC group (MCI: 2.09(0.34), NC: 1.87(0.35), p‐value: 0.024). After six months of social interaction, MCI and NC group were not significantly different (MCI: 1.81(0.47), NC: 1.95(0.47), p‐value: 0.280). Moreover, in MCI group, baseline’s MSN was significantly higher than M6’ (p‐value: 0.017), while, in NC group, there was no significant difference between baseline and M6 (p‐value: 0.465).ConclusionsWe showed that, through six months of social interaction, MCI group’s MSN reduces from significantly higher than NC group at the baseline to not significantly different from NC group at M6. This indicates that the high dosage of interactive conversations reduces MSN in the MCI group.

UP-Net: Uncertainty-Supervised Parallel Network for Image Manipulation Localization

Image manipulation localization remains a hot topic due to its inherent semantic-independent nature and realistic needs. Virtually all localization studies are devoted to solving arbitrary tampering using multi-branch networks based on deep features or skip-connection structures based on full features, which may induce the loss of manipulation details or noisy interference from image semantics. This poses a challenge for existing localization methods to fully capture invisible manipulations, especially in post-processing settings and across dataset scenarios. To address the above issues, we propose an uncertainty-supervised parallel network (UP-Net) for image tampering localization that preserves more manipulation details while avoiding semantic noise. UP-Net cascades the frequency and RGB domains of the manipulated image as dual-domain embedding, instead of dual-domain parallel learning as in previous work. To learn semantic-independent manipulation features, two structurally identical parallel branches are designed to learn tampering inconsistencies from intermediate and deep coding features for gradually obtaining the initial and final localization predictions. Where attention-guided partial decoder (AGPD) integrates more precise manipulation edges and manipulation semantics without introducing additional noise by focusing on channel correlation and spatial dependence, making a significant contribution to performance. Moreover, the new concept of uncertainty-constrained loss supervision is introduced to guide UP-Net to continuously improve confidence in locating difficult pixels, which are easily misclassified due to post-processing operations. Experiments on three public manipulation datasets and two real challenge datasets show that our end-to-end UP-Net achieves significant performance in manipulation localization, generalization across datasets, and robustness compared to state-of-the-art methods.

Twin Pseudo-training for semi-supervised semantic segmentation

Recent advancements in semi-supervised semantic segmentation have demonstrated the effectiveness of utilizing pseudo labels supervision to mitigate the limitations of pixel-wise annotations. However, pseudo-labels generated using self-training techniques typically contain a significant amount of noise, which can impede the training process of the supervised model. In this study, we identify low- and high-level semantic errors as the two key factors that hinder the accuracy of pseudo labels. To fully exploit the potential of pseudo labels, we introduce a novel semi-supervised framework named Twin Pseudo-training (TPseudo), which employs a consistency and disagreement collaboration strategy. Specifically, we correct pseudo labels with a False-positive Filter (FPF) to reduce high-level semantic noise and refine low-level semantic biases using a Semantic Error Detector (SED). Lastly, we design a Self-Adaptive Weight (SAW) loss function based on a disagreement between two predictions to exploit each pixel of pseudo labels. Experimental results on the standard benchmarks PASCAL VOC2012 and Cityscapes demonstrate the efficacy of the proposed method.

Representing noise: stacked plots and the contrasting diplomatic ambitions of radio astronomy and post-punk.

Sketched in 1979 by graphic designer Peter Saville, the record sleeve of Joy Division's Unknown Pleasures seemingly popularized one of the most celebrated radio-astronomical images: the 'stacked plot' of radio signals from a pulsar. However, the sleeve's designer did not have this promotion in mind. Instead, he deliberately muddled the message it originally conveyed in a typical post-punk act of artistic sabotage. In reconstructing the historical events associated with this subversive effort, this essay explores how, after its adoption as an imaging device utilized in radio astronomy, the stacked plot gave representation to the diplomacy agendas of two distinct groups. The post-punk reworking of the stacked plot exemplified the ambition of this artistic movement to attack the images associated with social conventions and norms by amplifying their 'semantic noise', and, in so doing, seeking to negotiate a social space for those sharing these subversive goals. Conversely, radio astronomers used the stacked plot to display the presence of interfering radio transmitters in the frequencies exclusively allocated to astronomical research, thus advocating the removal of this electronic noise in the context of international telecommunication negotiations. The article thus shows how the representation of different types of noise through similar images shaped contrasting ambitions in the separate domains of science diplomacy and everyday diplomacy.

Semantic Noise and Conceptual Stagnation in Natural Language Processing

Semantic noise, the effect ensuing from the denotative and thus functional variability exhibited by different terms in different contexts, is a common concern in natural language processing (NLP). While unarguably problematic in specific applications (e.g., certain translation tasks), the main argument of this paper is that failing to observe this linguistic matter of fact as a generative effect rather than as an obstacle, leads to actual obstacles in instances where language model outputs are presented as neutral. Given that a common and long-standing challenge in NLP is the interpretation of ambiguous – i.e., semantically noisy – cases, this article focuses on an exemplar ambiguity-resolution task in NLP: the problem of anaphora in Winograd schemas. The main question considered is: to what extent is the standard approach to disambiguation in NLP subject to a stagnant “image of language”? And, can a transdisciplinary, dynamic approach combining linguistics and philosophy elucidate new perspectives on these possible conceptual shortcomings? In order to answer these questions we explore the term and concept of noise, particularly in its presentation as semantic noise. Owing to its definitional plurality, and sometimes even desirable unspecificity, the term noise is thus used as proof of concept for semantic generativity being an inherent characteristic in linguistic representation, and its concept is used to interrogate assumptions admitted in the resolution of Winograd schemas. The argument is speculative and theoretical in method, and the result is an analysis which provides an account of the fundamentally dialogical and necessarily open-ended effects of semantic noise in natural language.

Semantic noise in the Winograd Schema Challenge of pronoun disambiguation

The Winograd Schema Challenge (WSC) of pronoun disambiguation is a Natural Language Processing (NLP) task designed to test to what extent the reading comprehension capabilities of language models (LMs) can be compared to those of human subjects. It is generally assumed across the NLP literature that human subjects are capable of resolving this task because of their acquired commonsense knowledge, thus setting a commonsense benchmark for LMs, one which has even been proposed as an alternative to the Turing test. In the context of complex natural language communications, Shannon and Weaver observed that the act of semantic interpretation is subject to semantic noise (Shannon and Weaver, 1964 (1949)). Semantic noise is a constraint that ensues from terms exhibiting variable interpretations across contexts, presenting a challenge to the resolution of tasks such as the WSC. However, the main argument of this paper is that rather than seeing semantic noise as a challenge to otherwise unambiguous communication, it can also be understood as a functional quality of natural language, given that it results in the conceptual negotiation of terms. Failing to theoretically attend to this linguistic matter of fact leads to unintended problems in instances where NLP applications are offered as unbiased or objectively applicable solutions. To address this, this article offers a renewed and original analysis of a series of Winograd Schemas, in order to demonstrate how they are not as straightforwardly solvable by human subjects as is commonly claimed across the NLP literature. The methodology employed is that of historical contextualisation in information theory, and qualitative cultural analysis drawing on examples from a wide variety of recent NLP literature.

A Semantic-Enhancement-Based Social Network User-Alignment Algorithm

User alignment can associate multiple social network accounts of the same user. It has important research implications. However, the same user has various behaviors and friends across different social networks. This will affect the accuracy of user alignment. In this paper, we aim to improve the accuracy of user alignment by reducing the semantic gap between the same user in different social networks. Therefore, we propose a semantically enhanced social network user alignment algorithm (SENUA). The algorithm performs user alignment based on user attributes, user-generated contents (UGCs), and user check-ins. The interference of local semantic noise can be reduced by mining the user's semantic features for these three factors. In addition, we improve the algorithm's adaptability to noise by multi-view graph-data augmentation. Too much similarity of non-aligned users can have a large negative impact on the user-alignment effect. Therefore, we optimize the embedding vectors based on multi-headed graph attention networks and multi-view contrastive learning. This can enhance the similar semantic features of the aligned users. Experimental results show that SENUA has an average improvement of 6.27% over the baseline method at hit-precision30. This shows that semantic enhancement can effectively improve user alignment.

Multi-label weak-label learning via semantic reconstruction and label correlations

In the multi-label classification task, an instance is simultaneously associated with multiple semantic labels. Due to the high complexity of the semantic space in practical applications, obtaining instances with full labels is difficult, leading to the weak-label problem. Existing methods focus on the low-rank and instance manifold regularization assumptions of the label matrix to recover the ground-truth label matrix but ignore the influence that the above assumptions may not hold due to the semantic noise caused by missing labels. To address the problem, this paper proposes a simple and effective method to recover the label space by reconstructing the label semantic space through joint label correlation to solve the multi-label weak-label classification task. Specifically, we leverage the label information consistency and feature-label dependency assumptions to reconstruct the semantic space and consider label correlations to enhance the information of semantic views. Moreover, l2,1-norm is utilized to mitigate the effect of missing label space noise. Additionally, the linear model of the proposed method is expanded to a nonlinear version of the kernel method to address the problem of the inseparability of linear data. Extensive experiments on several real-world tasks show that the proposed method outperforms some state-of-the-art methods.

Improving cross-lingual text matching with dual-level collaborative coarse-to-fine filter alignment network

Semantic alignment is a key component in Cross-Language Text Matching (CLTM) to facilitate matching (e.g., query-document matching) between two languages. The current solutions for semantic alignment mainly perform word-level translation directly, without considering the contextual information for the whole query and documents. To this end, we propose a Dual-Level Collaborative Rough-to-Fine Filter Alignment Network (DLCCFA) to achieve better cross-language semantic alignment and document matching. DLCCFA is devised with both a coarse-grained filter in word-level and a fine-grained filter in sentence-level. Concretely, for the query in word-level, we firstly extract top-k translation candidates for each token in the query through a probabilistic bilingual lexicon. Then, a Translation Probability Attention (TPA) mechanism is proposed to obtain coarse-grained word alignment, which generates the corresponding query auxiliary sentence. Afterwards, we further propose a Bilingual Cross Attention and utilize Self-Attention to achieve fine-grained sentence-level filtering, resulting in the cross-language representation of the query. The idea is that each token in the query works as an anchor to filter the semantic noise in the query auxiliary sentence and accurately align semantics of different languages. Extensive experiments on four real-world datasets of six languages demostrate that our method can outperform the mainstream alternatives of CLTM.

Three-Dimensional Human Head Reconstruction Using Smartphone-Based Close-Range Video Photogrammetry

Creation of head 3D models from videos or pictures of the head by using close-range photogrammetry techniques has many applications in clinical, commercial, industrial, artistic, and entertainment areas. This work aims to create a methodology for improving 3D head reconstruction, with a focus on using selfie videos as the data source. Then, using this methodology, we seek to propose changes for the general-purpose 3D reconstruction algorithm to improve the head reconstruction process. We define the improvement of the 3D head reconstruction as an increase of reconstruction quality (which is lowering reconstruction errors of the head and amount of semantic noise) and reduction of computational load. We proposed algorithm improvements that increase reconstruction quality by removing image backgrounds and by selecting diverse and high-quality frames. Algorithm modifications were evaluated on videos of the mannequin head. Evaluation results show that baseline reconstruction is improved 12 times due to the reduction of semantic noise and reconstruction errors of the head. The reduction of computational demand was achieved by reducing the frame number needed to process, reducing the number of image matches required to perform, reducing an average number of feature points in images, and still being able to provide the highest precision of the head reconstruction.

Discrete Semantics-Guided Asymmetric Hashing for Large-Scale Multimedia Retrieval

Cross-modal hashing technology is a key technology for real-time retrieval of large-scale multimedia data in real-world applications. Although the existing cross-modal hashing methods have achieved impressive accomplishment, there are still some limitations: (1) some cross-modal hashing methods do not make full consider the rich semantic information and noise information in labels, resulting in a large semantic gap, and (2) some cross-modal hashing methods adopt the relaxation-based or discrete cyclic coordinate descent algorithm to solve the discrete constraint problem, resulting in a large quantization error or time consumption. Therefore, in order to solve these limitations, in this paper, we propose a novel method, named Discrete Semantics-Guided Asymmetric Hashing (DSAH). Specifically, our proposed DSAH leverages both label information and similarity matrix to enhance the semantic information of the learned hash codes, and the ℓ2,1 norm is used to increase the sparsity of matrix to solve the problem of the inevitable noise and subjective factors in labels. Meanwhile, an asymmetric hash learning scheme is proposed to efficiently perform hash learning. In addition, a discrete optimization algorithm is proposed to fast solve the hash code directly and discretely. During the optimization process, the hash code learning and the hash function learning interact, i.e., the learned hash codes can guide the learning process of the hash function and the hash function can also guide the hash code generation simultaneously. Extensive experiments performed on two benchmark datasets highlight the superiority of DSAH over several state-of-the-art methods.

اللغة الإعلامية المعاصرة

الازدواجية | الخطاب الإعلامي | الحس اللغوي | الانهزامية | التواصل الاجتماعي | اللغة الوطنية | اللغة الرسمية | التشويش الدلالي | Double Standard | Informational Discourse | Contemporary Globalisation | Linguistic Sense | Defeatist | Social Media | The National Language Official Language | Semantic Noise

The gleam-glum effect: /i:/ versus /λ/ phonemes generically carry emotional valence.

The gleam-glum effect is a novel sound symbolic finding that words with the /i:/-phoneme (like gleam) are perceived more positive emotionally than matched words with the /Λ/-phoneme (like glum). We provide data that not only confirm the effect but also are consistent with an explanation that /i:/ and /Λ/ articulation tend to co-occur with activation of positive versus negative emotional facial musculature respectively. Three studies eliminate selection bias by including all applicable English words from the English Lexicon Project (Balota et al., 2007) and the Warriner et al. (2013) database and every possible Mandarin Pinyin combination that differ only in the middle phoneme (/i:/ vs /Λ/). In Study 1, 61 U.S. undergraduates rated monosyllabic English /i:/ words as robustly more positive than matched /Λ/ words. Study 2 analyzed the Warriner et al. (2013) valence ratings, extending the gleam-glum effect to all applicable words in the database. In Study 3, 38 U.S. participants (using English) and 37 participants in China (using Mandarin Pinyin) rated word pairs under three conditions that moderate musculature activity: Read aloud (Enhance), read silently (Control), and read silently while chewing gum (Interfere). Indeed, the effect was both replicated and was significantly larger when facial musculature was enhanced than when interfered with, and the two language populations did not significantly differ. These findings confirm a robust gleam-glum effect, despite semantic noise, in English and Mandarin Pinyin. Furthermore, these data are consistent with the hypothesis that this type of sound symbolism arises from the overlap in muscles used both in articulation and emotion expression. (PsycInfo Database Record (c) 2021 APA, all rights reserved).