Semantic Clues Research Articles

A classification approach to coreference in discharge summaries: 2011 i2b2 challenge

To create a highly accurate coreference system in discharge summaries for the 2011 i2b2 challenge. The coreference categories include Person, Problem, Treatment, and Test. An integrated coreference resolution system was developed by exploiting Person attributes, contextual semantic clues, and world knowledge. It includes three subsystems: Person coreference system based on three Person attributes, Problem/Treatment/Test system based on numerous contextual semantic extractors and world knowledge, and Pronoun system based on a multi-class support vector machine classifier. The three Person attributes are patient, relative and hospital personnel. Contextual semantic extractors include anatomy, position, medication, indicator, temporal, spatial, section, modifier, equipment, operation, and assertion. The world knowledge is extracted from external resources such as Wikipedia. Micro-averaged precision, recall and F-measure in MUC, BCubed and CEAF were used to evaluate results. The system achieved an overall micro-averaged precision, recall and F-measure of 0.906, 0.925, and 0.915, respectively, on test data (from four hospitals) released by the challenge organizers. It achieved a precision, recall and F-measure of 0.905, 0.920 and 0.913, respectively, on test data without Pittsburgh data. We ranked the first out of 20 competing teams. Among the four sub-tasks on Person, Problem, Treatment, and Test, the highest F-measure was seen for Person coreference. This system achieved encouraging results. The Person system can determine whether personal pronouns and proper names are coreferent or not. The Problem/Treatment/Test system benefits from both world knowledge in evaluating the similarity of two mentions and contextual semantic extractors in identifying semantic clues. The Pronoun system can automatically detect whether a Pronoun mention is coreferent to that of the other four types. This study demonstrates that it is feasible to accomplish the coreference task in discharge summaries.

Sur l'indice suggérant sémantique dans la perspective fonctionnelle de la phrase

In the Brno approach to the theory of Functional Sentence Perspective (FSP), four factors have been identified to work in mutual cooperation in rendering the final functional perspective of an utterance: actual linear arrangement of sentence elements, their dynamic semantic functions (the semantic factor), context, and – in spoken discourse – prosodic features. In a discussion of the semantic factor, Jan Firbas briefly describes in his monograph of 1992 the operation of the so-called suggestive semantic clue. Since this term has received little attention elsewhere, the author of the present paper attempts to show that it represents one of the fundamental concepts in FSP analysis in that it is not only closely attached to the important concept of semantic homogeneity, but also an FSP signal (or an FSP factor?) capable of contributing significantly to the correct resolution of the communicative functions of sentence elements.

Validity of the Five-Word Test for the Evaluation of Verbal Episodic Memory and Dementia in a Memory Clinic Setting

The five-word test (FWT) uses semantic clues to optimize the encoding and retrieval of 5 items. Our objective was to assess the validity of the FWT as a measure of episodic memory when compared with the Free and Cued Selective Reminding Test (FCSRT), and its ability to distinguish participants with any dementia and especially Alzheimer disease (AD) from those with only subjective complaints. Two hundred participants with Mini-Mental State Examination (MMSE) >15 were prospectively evaluated. The sum of the immediate and delayed free recalls of the FWT is called the free recall score (FRS). The sum of the immediate free, immediate cued, delayed free, and delayed cued recalls is called the total recall score (TRS). A total weighted score (TWS) is calculated using this formula: (free recalls × 2) + cued recalls. The correlation between FRS and the free recall scores of the FCSRT and between TRS and the total recall scores of the FCSRT was significant (r (s) ranges from .74-.84, P < .001). Area under the receiver--operating characteristic (ROC) curves of the MMSE, FRS, TRS, and TWS were not statistically different. A TWS at a cutoff value ≤15 could discriminate any dementia from subjective complaints with a sensitivity of 75% and a specificity of 95.9% or AD from subjective complaints with a sensitivity of 90.2% and a specificity of 95.9%. The FWT is a valid test of verbal episodic memory. It is useful to discriminate dementia especially AD from isolated subjective complaints.

Visual storylines: Semantic visualization of movie sequence

This paper presents a video summarization approach that automatically extracts and visualizes movie storylines in a static image for the purposes of efficient representation and quick overview. A new type of video visualization, Visual Storylines, is designed to summarize video storylines in an image composition while preserving the style of the original videos. This is achieved with a series of video analysis, image synthesis, relationship quantification and geometric layout optimization techniques. Specifically, we analyze the video contents and quantify the video story unit relationships automatically through clustering video shots according to both the visual and audio data. A multi-level storyline visualization method then organizes and synthesizes a suitable amount of representative information, including both the locations and interested objects and characters, with the assistance of arrows, according to the relationships between the video story units and the temporal structure of the video sequence. Several results have demonstrated that our approach is able to abstract the main storylines of professionally edited video such as commercial movies and TV series, though some semantic key clues might be missed in the summarization. Preliminary user studies have been performed to evaluate our approach, and the results show that our approach can be used to assist viewers to understand video contents when they are familiar with the context of the video or when a text synopsis is provided.

Preface for special issue on “Emotional intelligence and ambient intelligence”

In psychology, emotion that reflects the genuine inner feeling of a person experiencing encounters is shown to be a major index in the evaluation of cognition, behavior, and social skills. Emotion is also one source controlling the learning efficiency in education, and one of essential elements in realizing smart interaction for a computer. Thus, emotional intelligence, which is defined as the capability to be able to perceive, to assess and to manage emotions of one’s self or others emotion is attracting its attention in psychology cognition, education, and machine intelligence for achieving ambient intelligence. Understanding one’s emotion can be achieved from several observations, including facial expression, voice expression, and physiological signals. Emotion revealing is affected by culture and personal characteristics; therefore, detection/understanding human emotion is highly challenging. On the other hand, how and why emotion affects human mental states and reactions is still a mystery. All of these are major issues for realizing an emotional intelligence smart environment. The goal of this special issue is to provide a forum for bringing the experts from cross disciplinary to address the emerging topic of emotional intelligence. After rigorous reviewing and accurate revision, seven papers were selected for publication. Tan et al. applied two bipolar facial electromyography (EMG) channels over corrugator supercilii and zygomaticus for differentiating the emotional states visual stimuli in the valence arousal dimensions. Experimental results show that corrugator EMG and zygomaticus EMG efficiently differentiated negative and positive emotions. Lee et al. developed a regularized discriminant analysis (RDA)-based boosting algorithm, and applied it on the facial emotion recognition. The small sample size and ill-posed problems suffered from QDA and LDA was resolved in the paper through a regularization technique. They also used a particle swarm optimization (PSO) algorithm to estimate optimal parameters in RDA. Lin et al. constructed an extensible lexicon and use semantic clues to analyze the emotions of sentences posted on the Plurk website. A support vector machine is applied to classify the emotions. Cerezo et al. proposed a facial affect recognizer to sense emotions from users’ facial image. Five classifiers were integrated to identify emotions. In addition, a Kalman filtering technique was applied to ensure the temporal consistency and increase the robustness. Chi et al. investigates a clean train/noisy test scenario to simulate practical conditions with unknown noisy sources. They extracted statistics of joint spectro-temporal modulation features from an auditory perceptual model for the detection of the emotion status of the speech samples which are corrupted with white and babble noise under various SNR levels. Ana et al. built a virtual pet by using the Ortony, Clore and Collins’ (OCC) theory to implement a cognitive structure. The methodology starts from developing a Behavior Cognitive Task Analysis (BCTA) to elucidate the components necessary to simulate behaviors and mental models of virtual pets. In particular, the Fuzzy C-Mean (FCM) is also proposed to map interaction between elements in the emotion model. Fu et al. proposed a TVP.-C. Chung (&) National Cheng Kung University, Taiwan, R.O.C. e-mail: pcchung@eembox.ee.ncku.edu.tw

An emotion recognition mechanism based on the combination of mutual information and semantic clues

Recently, text-based emotion recognition has become more and more highly regarded. After exploring the related research, we have found that: (1) In the semantics-based solution, the newly emerging compound words cannot be handled. (2) In the statistics-based solution, the emotional values are computed only via the linear text clues; it lacks syntactic and semantic structural information. Therefore, in our effort to overcome the two concerns above, in the system presented here we construct an extensible lexicon and use semantic clues to analyze the sentences. In this work, we collected the sentences posted to the Plurk website as our corpus. The emoticons are classified into four types based on Thayer’s 2-D Model which is composed of valence (positive/negative emotions) and arousal (the strength of emotions). The system will pre-process the sentence to eliminate the useless information, and then transform it to an emotion lexicon. Additionally, this research analyzes three kinds of semantic clues: negation, transition, and coordinating conjunctions. The final emotion is decided by SVM and the merging algorithm proposed in this work. Moreover, the recognition rate is promoted from 40.3 to 68.15% after the emotion lexicon is improved, and the semantic structural information is applied. The experimental results show that the work is promising.

Episodic and semantic memory in children with mesial temporal sclerosis

The aim of this study was to analyze semantic and episodic memory deficits in children with mesial temporal sclerosis (MTS) and their correlation with clinical epilepsy variables. For this purpose, 19 consecutive children and adolescents with MTS (8 to 16 years old) were evaluated and their performance on five episodic memory tests (short- and long-term memory and learning) and four semantic memory tests was compared with that of 28 healthy volunteers. Patients performed worse on tests of immediate and delayed verbal episodic memory, visual episodic memory, verbal and visual learning, mental scanning for semantic clues, object naming, word definition, and repetition of sentences. Clinical variables such as early age at seizure onset, severity of epilepsy, and polytherapy impaired distinct types of memory. These data confirm that children with MTS have episodic memory deficits and add new information on semantic memory. The data also demonstrate that clinical variables contribute differently to episodic and semantic memory performance.

Modeling spatial and semantic cues for large-scale near-duplicated image retrieval

Bag-of-visual Words (BoW) image representation has been illustrated as one of the most promising solutions for large-scale near-duplicated image retrieval. However, the traditional visual vocabulary is created in an unsupervised way by clustering a large number of image local features. This is not ideal because it largely ignores the semantic and spatial contexts between local features. In this paper, we propose the geometric visual vocabulary which captures the spatial contexts by quantizing local features in bi-space, i.e., in descriptor space and orientation space. Then, we propose to capture the semantic context by learning a semantic-aware distance metric between local features, which could reasonably measure the semantic similarities between image patches, from which the local features are extracted. The learned distance is hence utilized to cluster the local features for semantic visual vocabulary generation. Finally, we combine the spatial and semantic contexts in a unified framework by extracting local feature groups, computing the spatial configurations between the local features inside the group, and learning a semantic-aware distance between groups. The learned group distance is then utilized to cluster the extracted local feature groups to generate a novel visual vocabulary, i.e., the contextual visual vocabulary. The proposed visual vocabularies, i.e., geometric visual vocabulary, semantic visual vocabulary and contextual visual vocabulary are tested in large-scale near-duplicated image retrieval applications. The geometric visual vocabulary and semantic visual vocabulary achieve better performance than the traditional visual vocabulary. Moreover, the contextual visual vocabulary, which combines both spatial and semantic clues outperforms the state-of-the-art bundled feature in both retrieval precision and efficiency.

Van boekenbonnen en feëverhale - De tussenklank e(n) in Nederlandse en Afrikaanse samenstellingen: vorm of betekenis?

Compounds in Dutch and Afrikaans may contain a linking sound resembling the plural suffix [ə] or [ən]. This linking sound can be spelled in Dutch as e or en, a difference which previous research has shown to affect the interpretation of left hand members of compounds regarding their plurality. Here, we look at the existence of the reverse effect: do semantic clues about number influence preferences for a linking sound in novel compounds? Indeed, contexts in which the left hand member of a compound-to-be is present in its plural form increase the tendency to prefer a linking element (using a seven-point rating scale). In contexts which merely suggest a plural interpretation for the left hand member, this propensity is markedly less for speakers of Dutch, and absent for speakers of Afrikaans. Of interest is the asymmetry with previous research that shows how subtle form differences influence number interpretation.

The Role of Clues in Chinese Idiom Riddle Solving

Chinese riddle solving is one of the insight problems in which Chinese idioms frequently turn out to be the answers. This study was to investigate the role of clues used by problem solvers when given such riddles. Three experiments were conducted and idiom riddles were used as experimental materials. In all experiments, a 2 (semantic clue vs. rule clue)×2 (validity vs. invalidity) experimental design was adopted. In Experiment 1, all clues were presented above visual threshold. Participants were asked to figure out answers themselves according to each riddle and its clue. In Experiment 2, except that the clues were presented below visual threshold, the others was the same as that in Experiment 1. In both of these two experiments, the participants were first required to judge the emotional value of 160 idioms within 5 minutes, which include the answers of the riddles used in the experiments. In Experiment 3, participants were asked to make a choice among six options. There were 206 first-year and second-year college students participated in these experiments. The results were as follow: (1) The valid rule clues, which were presented either above threshold or below threshold, facilitated the process of Chinese idiom riddle solving. (2) The response accuracy was increased when the semantic clues were valid. Moreover, the clues presented above the threshold played a bigger role than those presented below the threshold. (3) An interference was found only when the invalid clues were presented above the threshold. Based on the theories and hypotheses about insight problem solving, we put forward a model to explain the process of Chinese riddle solving, which include three phases.

Investigating readers’ mental maps of references in an online system

Referential identification and resolution are considered the keys to help readers grasp the main idea of a text and solve lexical ambiguities. The goal of this study is to design a computer system for helping college students who learn English as a Foreign Language (EFL) develop mental maps of referential identification and resolution in reading. Four modules, Natural Language Processing (NLP), User Interface, Recording, and Feedback Tool, are implemented in the system. Results of this study showed that the more-proficient EFL readers were able to identify and resolve most of the references to form a coherent mental map from different parts of a text. The less-proficient readers commonly resolved references by relying on grammatical rules instead of semantic contextual clues. They often referred references to incorrect objects. To overcome the difficulties in figuring out the relationship between two words, the less-proficient readers usually asked for more feedbacks. As students progressed in reading, they requested fewer feedbacks in the online system. Some recommendations for future studies are discussed.

A New Approach for Overlay Text Detection and Extraction From Complex Video Scene

Overlay text brings important semantic clues in video content analysis such as video information retrieval and summarization, since the content of the scene or the editor's intention can be well represented by using inserted text. Most of the previous approaches to extracting overlay text from videos are based on low-level features, such as edge, color, and texture information. However, existing methods experience difficulties in handling texts with various contrasts or inserted in a complex background. In this paper, we propose a novel framework to detect and extract the overlay text from the video scene. Based on our observation that there exist transient colors between inserted text and its adjacent background, a transition map is first generated. Then candidate regions are extracted by a reshaping method and the overlay text regions are determined based on the occurrence of overlay text in each candidate. The detected overlay text regions are localized accurately using the projection of overlay text pixels in the transition map and the text extraction is finally conducted. The proposed method is robust to different character size, position, contrast, and color. It is also language independent. Overlay text region update between frames is also employed to reduce the processing time. Experiments are performed on diverse videos to confirm the efficiency of the proposed method.

Content-Based 3-D Model Retrieval: A Survey

As the number of available 3D models grows, there is an increasing need to index and retrieve them according to their contents. This paper provides a survey of the up-to-date methods for content-based 3D model retrieval. First, the new challenges encountered in 3D model retrieval are discussed. Then, the system framework and some key techniques of content-based 3D model retrieval are identified and explained, including canonical coordinate normalization and preprocessing, feature extraction, similarity match, query representation and user interface, and performance evaluation. In particular, similarity measures using semantic clues and machine learning methods, as well as retrieval approaches using nonshape features, are given adequate recognition as improvements and complements for traditional shape-matching techniques. Typical 3D model retrieval systems and search engines are also listed and compared. Finally, future research directions are indicated, and an extensive bibliography is provided.

Latent linkage semantic kernels for collective classification of link data

Generally, links among objects demonstrate certain patterns and contain rich semantic clues. These important clues can be used to improve classification accuracy. However, many real-world link data may exhibit more complex regularity. For example, there may be some noisy links that carry no human editorial endorsement about semantic relationships. To effectively capture such regularity, this paper proposes latent linkage semantic kernels (LLSKs) by first introducing the linkage kernels to model the local and global dependency structure of a link graph and then applying the singular value decomposition (SVD) in the kernel-induced space. For the computational efficiency on large datasets, we also develop a block-based algorithm for LLSKs. A kernel-based contextual dependency network (KCDN) model is then presented to exploit the dependencies in a network of objects for collective classification. We provide experimental results demonstrating that the KCDN model, together with LLSKs, demonstrates relatively high robustness on the datasets with the complex link regularity, and the block-based computation method can scale well with varying sizes of the problem.

Personal Name Resolution Crossover Documents by a Semantics-Based Approach

Cross-document personal name resolution is the process of identifying whether or not a common personal name mentioned in different documents refers to the same individual. Most previous approaches usually rely on lexical matching such as the occurrence of common words surrounding the entity name to measure the similarity between documents, and then clusters the documents according to their referents. In spite of certain successes, measuring similarity based on lexical comparison sometimes ignores important linguistic phenomena at the semantic level such as synonym or paraphrase. This paper presents a semantics-based approach to the resolution of personal name crossover documents that can make the most of both lexical evidences and semantic clues. In our method, the similarity values between documents are determined by estimating the semantic relatedness between words. Further, the semantic labels attached to sentences allow us to highlight the common personal facts that are potentially available among documents. An evaluation on three web datasets demonstrates that our method achieves the better performance than the previous work.

Puzzling Thoughts for H. M.: Can New Semantic Information Be Anchored to Old Semantic Memories?

Researchers currently debate whether new semantic knowledge can be learned and retrieved despite extensive damage to medial temporal lobe (MTL) structures. The authors explored whether H. M., a patient with amnesia, could acquire new semantic information in the context of his lifelong hobby of solving crossword puzzles. First, H. M. was tested on a series of word-skills tests believed important in solving crosswords. He also completed 3 new crosswords: 1 puzzle testing pre-1953 knowledge, another testing post-1953 knowledge, and another combining the 2 by giving postoperative semantic clues for preoperative answers. From the results, the authors concluded that H. M. can acquire new semantic knowledge, at least temporarily, when he can anchor it to mental representations established preoperatively.

Hebrew Key-Letters

This article uses mostly ignored facts about Hebrew to propose a major addition to the theory of the Hebrew root (and root-morphemes generally), and a radical solution to the lexical problem in linguistics: the fact that native speakers of a language acquire and subsequently access tens of thousands of words (and that knowledge of a significant part of these words is required of foreign learners). Initial single-consonant submorphemes are proposed as a new (quasi-) systematic structure within lexicon, most clearly seen in Hebrew. More specifically, it is proposed that 16 Hebrew initial consonants provide a semantic clue to the meaning of the root, in the form of a small set of basic meanings (one "original" one branching out to several basic ones). These clues help to learn and access a large vocabulary, whether for the native or non-native speaker.

Les ambiguïtés

Summary Homography of utterances (words or meaningful sequences of words) is a term used for ambiguity at the spelling level. Ambiguity is the preferred term for phrases and sentences. The present paper is an attempt to clarify the status of this notion. When people talk, write, read or listen to discourses, they seldom raise questions of ambiguity of words or phrases. Discourses, spoken or written, are rarely perceived as ambivalent at the linguistic level, hence the words from which they are composed are not ambiguous. Questions of ambiguity arise immediately when words are isolated from any context. In such cases, the discussion of the meaning of a word amounts to defining its dictionary entry, that is to enumerating its different meanings. The discrepancy between the two situations is summed up in one sentence: Contexts resolve ambiguities. When a word, a priori ambiguous according to dictionaries, is put to use in a discourse, this discourse filters out all irrelevant interpretations of the word. The organization of a dictionary is intended to provide contexts for the use of words. This information takes on different forms: –definitions incorporate semantic clues of use, –examples of usage provide syntactic and semantic information. Such information cannot be considered as explicit contexts, thus it is up to the human user to generalize these indications in order to apply them to the interpretation of texts. We present a number of situations where contexts can be formalized and systematically described.

The Tambartun Oral Reading Test: A new test for determining reading performance of the visually impaired

A new Norwegian oral reading test has been developed and tested on a group of visually impaired subjects and a group of normally sighted controls. Twenty-eight equivalent reading test charts, each consisting of 50 words, were made. The 50 words in each chart were randomly selected from the 300 two- to six-letter words most frequently used in the Norwegian language. All words in the test charts are unrelated, which means that the reader cannot base his or her reading on syntactic and semantic clues available in a meaningful text. The functional equivalence of the elements upon which the test is built makes it possible to compare reading facility under different physical conditions, e.g. under different light levels.The test is simple and quick to perform. This paper reports on the test's reliability and validity and its possible application in reading rehabilitation of the visually impaired.

Syntactic and semantic processing in Hebrew readers with prelingual deafness.

The experimenter investigated the effect of semantic clues on the reading comprehension of deaf and hearing Israeli children. Two groups of students with prelingual deafness, and a hearing control group, were asked to read syntactically simple and syntactically relative sentences of varying semantic plausibility. Sixteen of the participants who were deaf (mean grade 6.9) had been trained orally, using spoken language as their principal means of communication at home and at school. Another 16 students with deafness (mean grade 6.9), all of them children of deaf parents, had acquired sign language as their primary language. The mean grade of the hearing control group was 6.5. The results suggest that, in contrast to the case with hearing individuals, reading comprehension in individuals with prelingually acquired deafness, regardless of communication background, is predominantly determined by the semantic processing of content words, with only minor attention given to the processing of the syntactic structure of the text.