Semantic Mismatch Research Articles

Content analysis-based documentation and exploration of research articles

PurposeWith the wealth of information available on the World Wide Web, it is difficult for anyone from a general user to the researcher to easily fulfill their information need. The main challenge is to categorize the documents systematically and also take into account more valuable data such as semantic information. The purpose of this paper is to develop a concept-based search system that leverages the external knowledge resources as the background knowledge for getting the accurate and efficient meaningful search results.Design/methodology/approachThe paper introduces the approach which is based on formal concept analysis (FCA) with the semantic information to support the document management in information retrieval (IR). To describe the semantic information of the documents, the system uses the popular knowledge resources WordNet and Wikipedia. By using FCA, the system creates the concept lattice as the concept hierarchy of the document and proposes the navigation algorithm for retrieving the hierarchy based on the user query.FindingsThe semantic information of the document is based on the two external popular knowledge resources; the authors find that it will be more efficient to deal with the semantic mismatch problems of user need.Originality/valueThe navigation algorithm proposed in this research is applied to the scientific articles of the National Science Foundation (NSF). The proposed system can enhance the integration and exploration of the scientific articles for the advancement of the Scientific and Engineering Research Community.

Cross-Layer Distillation with Semantic Calibration

Recently proposed knowledge distillation approaches based on feature-map transfer validate that intermediate layers of a teacher model can serve as effective targets for training a student model to obtain better generalization ability. Existing studies mainly focus on particular representation forms for knowledge transfer between manually specified pairs of teacher-student intermediate layers. However, semantics of intermediate layers may vary in different networks and manual association of layers might lead to negative regularization caused by semantic mismatch between certain teacher-student layer pairs. To address this problem, we propose Semantic Calibration for Cross-layer Knowledge Distillation (SemCKD), which automatically assigns proper target layers of the teacher model for each student layer with an attention mechanism. With a learned attention distribution, each student layer distills knowledge contained in multiple layers rather than a single fixed intermediate layer from the teacher model for appropriate cross-layer supervision in training. Consistent improvements over state-of-the-art approaches are observed in extensive experiments with various network architectures for teacher and student models, demonstrating the effectiveness and flexibility of the proposed attention based soft layer association mechanism for cross-layer distillation.

Controllable Text-to-Image Generation with Enhanced Text Encoder and Edge-Preserving Embedding

Text-to-image generation through generative adversarial networks (GANs) has become a popular research topic in the fields of natural language processing and high-quality image synthesizing. By learning latent word features of queries, many state-of-the-art GANs generate images from coarse to fine in a multi-stage trend. While this mechanism is able to synthesize realistic images stage by stage, failures containing semantic mismatches and shape distortions still exist. In this project, we adopted a multi-stage text-to-image GAN as a baseline, and remodeled this structure in two ways: 1) modify the text encoder by pre-processing text queries with pre-trained language models to gain deeply mined text information, and 2) develop an edge-preserving structure to capture feature mismatches, denoted as “edge loss”, between text latent information and generated image edges. Experiments on benchmark dataset demonstrate that models of our method outperform the baseline model and increase Inception Scores by over 30%, and that our approaches are able to effectively generate synthetic images using natural language descriptions.

Image Style Transfer Algorithm Based on Semantic Segmentation

Most of the existing image style transfer algorithms transfer the whole image style as a whole. Style feature is a set of correlation matrix based on style image, namely Gram matrix. Each matrix is a global description of the style image. This kind of methods can perform well in the insensitive semantic scenes (such as the style transfer between landscape photos), but in the sensitive semantic scenes (such as the style transfer between portrait photos), the problem of semantic mismatch will be highlighted, such as transferring the background texture of the style image to the foreground of the target image. Although the existing research takes the manually annotated semantic image as an input of the algorithm, and then guides the style transfer based on the semantic information, and finally achieves good results in the style transfer between portraits. But there are still two problems: first, semantic images need to be manually annotated, which costs human resources. In practical applications, large-scale image style transfer is often needed. Second, the details of the synthesized image are fuzzy, and the definition is not enough. We propose an image style transfer algorithm based on semantic segmentation to resolve semantic mismatching in image style transfer. Our algorithm extracts the semantic information of style image and content image automatically through a semantic segmentation network and uses the semantic information to guide the style transfer. Our algorithm builds a semantic segmentation network based on mask R-CNN, introduces semantic information, and then makes style transfer on the patch level, realizes the style transfer between similar objects (consistent semantic information). Experiments on Celeba and Wikiart show that our method could automatically extract the semantic information of style image and content image. Compared with the state-of-art approaches in this field, our method can effectively avoid semantic mismatch in the process of image style transfer. That is, it can maintain semantic consistency in the process of style transfer.

Semantic impairment in very early Alzheimer disease

AbstractBackgroundComparing to episodic memory, semantic memory is less likely to be impaired in normal aging. Investigation of semantic memory may give us a better chance on detecting Alzheimer disease in early stage.MethodIn 2019, we collected 20 patients from our memory clinic. Half of them were diagnosed as amnesic MCI (mild cognitive impairment), half of them were their spouse. All of them received amyloid PET (florbetaben), brain MRI, ERP (evoked related potential) studies and detailed language evaluation. According to their clinical symptoms and status of amyloid deposition, they were grouping as normal, preclinical and prodromal stage.ResultIn picture description test, participants of prodromal group expressed fewer semantic contents and more filler words compared to others. No significant difference was found between preclinical and normal group. However, N400 (component of ERP, indicating semantic mismatch) showed difference between preclinical and normal group.ConclusionSemantic network impairment begins from preclinical stage of AD. Through detailed spoken language analysis, we could use it as early biomarker and trace disease progression.

Learning Unsupervised Knowledge-Enhanced Representations to Reduce the Semantic Gap in Information Retrieval

The semantic mismatch between query and document terms—i.e., the semantic gap—is a long-standing problem in Information Retrieval (IR). Two main linguistic features related to the semantic gap that can be exploited to improve retrieval are synonymy and polysemy. Recent works integrate knowledge from curated external resources into the learning process of neural language models to reduce the effect of the semantic gap. However, these knowledge-enhanced language models have been used in IR mostly for re-ranking and not directly for document retrieval. We propose the Semantic-Aware Neural Framework for IR (SAFIR), an unsupervised knowledge-enhanced neural framework explicitly tailored for IR. SAFIR jointly learns word, concept, and document representations from scratch. The learned representations encode both polysemy and synonymy to address the semantic gap. SAFIR can be employed in any domain where external knowledge resources are available. We investigate its application in the medical domain where the semantic gap is prominent and there are many specialized and manually curated knowledge resources. The evaluation on shared test collections for medical literature retrieval shows the effectiveness of SAFIR in terms of retrieving and ranking relevant documents most affected by the semantic gap.

INTERLANGUAGE PARALLELS IN RUSSIAN AND BULGARIAN LEGAL TERMINOLOGY

Formally identical terminological units belonging to related languages can demonstrate both coincidence and divergence in semantics and stylistic correlation - due to linguistic reasons, as well as extralinguistic factors generated by the legal processes of a particular state. All this complicates the legal translation procedure, which means that it requires a detailed legal and linguistic analysis of such units.
 The article provides rationale for the systematization and comprehensive description of formally matching interlanguage correspondences of Slavic and non-Slavic origin, in which at least one lexical unit belongs to the legal terminology of the Russian or Bulgarian language. The terminological pairs with the same plan of expression identified and equally expressed in the modern Russian and Bulgarian legislation are placed on a scale of semantic correspondences as follows: zero (false) semantic equivalence - partial semantic similarity - full semantic equivalence.
 It is noted that particular difficulties can be correlated with the translation of formally identical terms-derived words that arose from the same elements based on the same word-formation model, as this makes their semantic mismatch not so obvious. The analysis of specific examples demonstrates that in closely related languages different units of parallel intralinguistic synonymic series can be codified, and therefore ignoring aspects of the word's functioning in the language leads to lexical and semantic interference in translation.
 The author concludes that a comparative study of interlanguage parallels in Russian and Bulgarian legal terminology is important for solving problems of lexicography and legal translation.

Effect of Speech Degradation and Listening Effort in Reverberating and Noisy Environments Given N400 Responses.

Background and ObjectivesIn distracting listening conditions, individuals need to pay extra attention to selectively listen to the target sounds. To investigate the amount of listening effort required in reverberating and noisy backgrounds, a semantic mismatch was examined.Subjects and MethodsElectroencephalography was performed in 18 voluntary healthy participants using a 64-channel system to obtain N400 latencies. They were asked to listen to sounds and see letters in 2 reverberated×2 noisy paradigms (i.e., Q-0 ms, Q-2000 ms, 3 dB-0 ms, and 3 dB-2000 ms). With auditory-visual pairings, the participants were required to answer whether the auditory primes and letter targets did or did not match.ResultsQ-0 ms revealed the shortest N400 latency, whereas the latency was significantly increased at 3 dB-2000 ms. Further, Q-2000 ms showed approximately a 47 ms delayed latency compared to 3 dB-0 ms. Interestingly, the presence of reverberation significantly increased N400 latencies. Under the distracting conditions, both noise and reverberation involved stronger frontal activation.ConclusionsThe current distracting listening conditions could interrupt the semantic mismatch processing in the brain. The presence of reverberation, specifically a 2000 ms delay, necessitates additional mental effort, as evidenced in the delayed N400 latency and the involvement of the frontal sources in this study.

A Nativity Sermon in the 13th Century Tolstovskiy Sbornik: Textology and Language Features

The paper focuses on the composition, lexical, and grammatical features of a Nativity sermon in the 13 th century Old Russian Tolstovskiy Sbornik (National Library of Russia, F.p.I.39). The author considers its Byzantine sources, principles of editorial work, and the differences from original rhetorical structures. Attributed to John Chrysostom, the sermon turns out to be a complicated compilation from various early Byzantine sermons. The compilation is based both on rearranging fragments of the same source and on combining excerpts from different sermons in a small context. Such transformations indicate the lack of cohesion in sermon texts, due to their independence from the causation and time factor. Non-attributed parts of the Old Russian text may be original since they demonstrate a certain similarity with Kirill Turovskiy orations in the same anthology. The lexical level of the sermon contains non-standard solutions that reinterpret the Greek source text, which may indicate either the missionary nature of the translation or a tendency to the poetic decoration. In some cases, the semantic mismatch of lexical units within Greek-Slavonic correlations is due to errors. At the grammatical level, there are also grammatical inconsistencies of Slavonic and Greek units; they affect the categories of time, number, gender, as well as parts-of-speech status.

SEPTIC: Detecting Injection Attacks and Vulnerabilities Inside the DBMS

Databases continue to be the most commonly used backend storage in enterprises, but they are often integrated with vulnerable applications, such as web frontends, which allow injection attacks to be performed. The effectiveness of such attacks stems from a semantic mismatch between how SQL queries are believed to be executed and the actual way in which databases process them. This leads to subtle vulnerabilities in the way input validation is done in applications. In this paper, we propose SEPTIC, a mechanism for DBMS attack prevention, which can also assist on the identification of the vulnerabilities in the applications. The mechanism was implemented in MySQL and evaluated experimentally with various applications and alternative protection approaches. Our results show no false negatives and no false positives with SEPTIC, on the contrary to other solutions. They also show that SEPTIC introduces a low performance overhead, in the order of 2.2%.

An Efficient LSM-Tree-Based SQLite-Like Database Engine for Mobile Devices

SQLite has been deployed in millions of mobile devices from Web to smartphone applications on various mobile operating systems. However, due to the uncoordinated IO interactions with the underlying file system (e.g., ext4), SQLite is not efficient with low transactions per second. In this paper, we for the first time propose a new SQLite-like database engine, called SQLiteKV, which adopts the LSM-tree-based data structure but retains the SQLite operation interfaces. With its SQLite interface, SQLiteKV can be utilized by existing applications without any modification, while providing high performance with its LSM-tree-based data structure. We separate SQLiteKV into the front-end and back-end. In the front-end, we develop a light-weight SQLite-to-KV compiler to solve the semantic mismatch, so SQL statements can be efficiently translated into KV operations. We also design a novel coordination caching mechanism with memory defragmentation so query results can be effectively cached inside SQLiteKV by alleviating the discrepancy of data management between front-end SQLite statements and back-end data organization. In the back-end, we adopt an LSM-tree-based key-value database engine, and propose a lightweight metadata management scheme to mitigate the memory requirement. We have implemented and deployed SQLiteKV on a Google Nexus 6P smartphone. Experimental results with various workloads show that SQLiteKV outperforms SQLite up to six times.

Auditory Distraction in Short-term Memory: Stable Effects of Semantic Mismatches on Serial Recall

ABSTRACTRecently, it was demonstrated that heard sentences containing semantically unexpected words disrupt visual-verbal serial recall more than sentences containing semantically expected words. This semantic mismatch effect did not become smaller over the course of the experiment, contrary to what has been observed with other semantic effects. This surprising finding was critically examined in the present study. In Experiment 1, unspecific habituation was investigated using a classical design and a larger number of trials compared to the original study. In Experiment 2, unspecific and specific habituation were investigated by presenting sixteen different distractor sentences in one condition and the same sentences sixteen times in the other condition. In both experiments, there was no evidence of habituation of the semantic mismatch effect. In Experiment 2, overall performance was significantly better with repeated distractor sentences as compared to different sentences, but the semantic mismatch effect remained unchanged. The disruptive effect of semantic mismatches on serial recall seems to be relatively resistant to habituation, suggesting a stable mechanism that allows to detect, and to react to, potentially meaningful information in the unattended channel.

Deep learning models and datasets for aspect term sentiment classification: Implementing holistic recurrent attention on target-dependent memories

An essential challenge in aspect term sentiment classification using deep learning is modeling a tailor-made sentence representation towards given aspect terms to enhance the classification performance. To seek a solution to this, we have two main research questions: (1) Which factors are vital for a sentiment classifier? (2) How will these factors interact with dataset characteristics? Regarding the first question, harmonious combination of location attention and content attention may be crucial to alleviate semantic mismatch problem between aspect terms and opinion words. However, location attention does not reflect the fact that critical opinion words usually come left or right of corresponding aspect terms, as implied in the target-dependent method although not well elucidated before. Besides, content attention needs to be sophisticated to combine multiple attention outcomes nonlinearly and consider the entire context to address complicated sentences. We merge all these significant factors for the first time, and design two models differing a little in the implementation of a few factors. Concerning the second question, we suggest a new multifaceted view on the dataset beyond the current tendency to be somewhat indifferent to the dataset in pursuit of a universal best performer. We then observe the interaction between factors of model architecture and dimensions of dataset characteristics. Experimental results show that our models achieve state-of-the-art or comparable performances and that there exist some useful relationships such as superior performance of bi-directional LSTM over one-directional LSTM for sentences containing multiple aspects and vice versa for sentences containing only one aspect.

Eliciting ERP Components for Morphosyntactic Agreement Mismatches in Perfectly Grammatical Sentences.

The present event-related brain potential (ERP) study investigates mechanisms underlying the processing of morphosyntactic information during real-time auditory sentence comprehension in French. Employing an auditory-visual sentence-picture matching paradigm, we investigated two types of anomalies using entirely grammatical auditory stimuli: (i) semantic mismatches between visually presented actions and spoken verbs, and (ii) number mismatches between visually presented agents and corresponding morphosyntactic number markers in the spoken sentences (determiners, pronouns in liaison contexts, and verb-final “inflection”). We varied the type and amount of number cues available in each sentence using two manipulations. First, we manipulated the verb type, by using verbs whose number cue was audible through subject (clitic) pronoun liaison (liaison verbs) as well as verbs whose number cue was audible on the verb ending (consonant-final verbs). Second, we manipulated the pre-verbal context: each sentence was preceded either by a neutral context providing no number cue, or by a subject noun phrase containing a subject number cue on the determiner. Twenty-two French-speaking adults participated in the experiment. While sentence judgment accuracy was high, participants' ERP responses were modulated by the type of mismatch encountered. Lexico-semantic mismatches on the verb elicited the expected N400 and additional negativities. Determiner number mismatches elicited early anterior negativities, N400s and P600s. Verb number mismatches elicited biphasic N400-P600 patterns. However, pronoun + verb liaison mismatches yielded this pattern only in the plural, while consonant-final changes did so in the singular and the plural. Furthermore, an additional sustained frontal negativity was observed in two of the four verb mismatch conditions: plural liaison and singular consonant-final forms. This study highlights the different contributions of number cues in oral language processing and is the first to investigate whether auditory-visual mismatches can elicit errors reminiscent of outright grammatical errors. Our results emphasize that neurocognitive mechanisms underlying number agreement in French are modulated by the type of cue that is used to identify auditory-visual mismatches.

A 3D assembly model retrieval method based on assembly information

PurposeThe purpose of this paper is to propose a three-dimensional (3D) assembly model retrieval method based on assembling semantic information to address semantic mismatches, poor accuracy and low efficiency in existing 3D assembly model retrieval methods.Design/methodology/approachThe paper proposes an assembly model retrieval method. First, assembly information retrieval is performed, and 3D models that conform to the design intention of the assembly are found by retrieving the code. On this basis, because there are conjugate subgraphs between attributed adjacency graphs (AAG) that have an assembly relationship, the assembly model geometric retrieval is translated into a problem of finding AAGs with a conjugate subgraph. Finally, the frequent subgraph mining method is used to retrieve AAGs with conjugate subgraphs.FindingsThe method improved the efficiency and accuracy of assembly model retrieval.Practical implicationsThe examples illustrate the specific retrieval process and verify the feasibility and reasonability of the assembly model retrieval method in practical applications.Originality/valueThe assembly model retrieval method in the paper is an original method. Compared with other methods, good results were obtained.

Prediction mechanisms in motor and auditory areas and their role in sound perception and language understanding

Is the meaning of an expected stimulus manifest in brain activity even before it appears? Although theories of predictive coding see anticipatory activity as crucial for the understanding of brain function, few studies have explored neurophysiologically manifest semantic predictions. Here, we report predictive negative-going potentials before the onset of action (i.e. whistle and hand clap) and non-action (i.e. pure tone, water drop) sounds. These prediction potentials (PP) indexed the meaning of action-related sounds. Dependent on the body-part-relationship of sound stimuli, neural sources were relatively more prominent in dorsal or ventral motor areas. In contrast, meaningless sounds (pure tones) activated predictive sources in temporal areas close to the auditory cortex; complex environmental sounds induced an anticipatory positivity broadly distributed over the scalp. We also found a systematic relationship between predictive activity and a Mismatch Negativity (MMN) like response to unexpected meaningful words which were presented as rare deviant stimuli amongst frequently repeated sounds. This deviant-elicited potential indexed semantic priming between action sounds and action-related words and semantic mismatch (prediction error). These results suggest a systematic link between perceptual/semantic prediction and matching mechanisms in the processing of sounds and words.

Semantic-aware neural style transfer

This study proposes a semantic-aware style transfer method for resolving semantic mismatch problems in existing algorithms. As the primary focus of this study, the consideration of semantic matching is expected to improve the quality of artistic style transfer. Here, each image is partitioned into several semantic regions for both a target photograph and a source painting. All partitioned regions of the target are then associated with one of the partitioned regions in the source according to their semantic interpretation. Given a pair of target and source regions, style is learned from the source region whereas content is learned from the target region. By integrating both the style and content components, we can successfully generate a stylized output. Unlike previous approaches, we obtain the best semantic match between regions using word embeddings. Thus, we guarantee that semantic matching is always established between the target and source. Moreover, it is unreliable to partition a painting using existing algorithms because of statistical gaps between the real photographs and paintings. To bridge such gaps, we apply a domain adaptation technique on the source painting to extract its semantic regions. We evaluated the effectiveness of the proposed algorithm based on a thorough experimental analysis and comparison. Through a user study, it is confirmed that semantic information considerably influences the quality assessment of style transfer.

A semantic mismatch effect on serial recall: Evidence for interlexical processing of irrelevant speech.

Short-term memory (STM) for serially presented visual items is disrupted by task-irrelevant, to-be-ignored speech. Five experiments investigated the extent to which irrelevant speech is processed semantically by contrasting the following two hypotheses: (1) semantic processing of irrelevant speech is limited and does not interfere with serial STM or (2) irrelevant speech is routinely processed semantically at an interlexical level to allow for the detection of stimuli that are of potential relevance for the individual. We tested these hypotheses by comparing the disruption of serial recall by distractor sentences with semantically expected endings to that of sentences with semantically unexpected endings. Sentences with unexpected endings consistently produced more disruption than sentences with expected endings. Phonologically expected, but semantically unexpected endings had the same effect, demonstrating that a semantic and not a phonological mismatch is responsible for the increased disruption. In all five experiments, the semantic mismatch effect was not reduced after repeated exposure to sentences with unexpected endings. The results suggest that (1) semantic processing occurs at the interlexical level of irrelevant speech, which requires the integration of words into the sentence context and (2) the semantic content of a distractor interferes with the maintenance of serial information in short-term-memory if it is unexpected. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

漢韓中動結構對比硏究

Middle Construction is a structure between initiative and passivity. Its subject is non agent component, while predicate verb is active form. Although the Middle Construction is active in form, it actually expresses passive meaning. The syntactic and semantic mismatch of Middle Constructions has attracted much attention from various linguists, and has been a hot topic in grammar research since the 80s of last century. The field of academic research mainly focuses on the formation of the middle dynamic structure, the cognitive motivation, the syntactic and semantic features, and the construction and sentence conditions of the middle verb, but some of the problems have not reached a consensus until now. The middle dynamic structure is the syntactic realization of the middle verb in different languages. It is also a universal and cross linguistic phenomenon. The study of the language phenomenon is of typological significance.BR Chinese shows a more obvious view based on the object, while Korean shows a more obvious view based on the main body. Because the Middle Construction is the weakening of the practice, the difference between the cognitive perspective and the cognitive perspective determines the dynamic structure of the Korean language, the Chinese Middle construction is more productive, the content of the expression can be more rich, and the applicable context is more. There are similarities and differences between Chinese and Korean Middle Constructions. Their differences are determined by the different linguistic attributes of Chinese and Korean.

Video summarization using event‐related potential responses to shot boundaries in real‐time video watching

Our aim was to develop an event‐related potential (ERP)‐based method to construct a video skim consisting of key shots to bridge the semantic gap between the topic inferred from a whole video and that from its summary. Mayer's cognitive model was examined, wherein the topic integration process of a user evoked by a visual stimulus can be associated with long‐latency ERP components. We determined that long‐latency ERP components are suitable for measuring a user's neuronal response through a literature review. We hypothesized that N300 is specific to the categorization of all shots regardless of topic relevance, N400 is specific for the semantic mismatching process for topic‐irrelevant shots, and P600 is specific for the context updating process for topic‐relevant shots. In our experiment, the N400 component led to more negative ERP signals in response to topic‐irrelevant shots than to topic‐relevant shots and showed a fronto‐central scalp pattern. P600 elicited more positive ERP signals for topic‐relevant shots than for topic‐irrelevant shots and showed a fronto‐central scalp pattern. We used discriminant and artificial neural network (ANN) analyses to decode video shot relevance and observed that the ANN produced particularly high success rates: 91.3% from the training set and 100% from the test set.