Character Levels Research Articles

Transliteration of Arabizi into Arabic Script for Tunisian Dialect

The evolution of information and communication technology has markedly influenced communication between correspondents. This evolution has facilitated the transmission of information and has engendered new forms of written communication (email, chat, SMS, comments, etc.). Most of these messages and comments are written in Latin script, also called Arabizi . Moreover, the language used in social media and SMS messaging is characterized by the use of informal and non-standard vocabulary, such as repeated letters for emphasis, typos, non-standard abbreviations, and nonlinguistic content like emoticons. Since the Tunisian dialect suffers from the unavailability of basic tools and linguistic resources compared to Modern Standard Arabic, we resort to the use of these written sources as a starting point to build large corpora automatically. In the context of natural language processing and to benefit from these networks’ data, transliterating from Arabizi to Arabic script is a necessary step because most recently available tools for processing the Tunisian dialect expect Arabic script input. Indeed, the transliteration task can help construct and enrich parallel corpora and dictionaries for the Tunisian dialect and can be useful for developing various natural language processing applications such as sentiment analysis, opinion mining, topic detection, and machine translation. In this article, we focus on converting the Tunisian dialect text that is written in Latin script to Arabic script following the Conventional Orthography for Dialectal Arabic. Then, we propose two models to transliterate Arabizi into Arabic script for the Tunisian dialect, namely a rule-based model and a discriminative model as a sequence classification task based on conditional random fields). In the first model, we use a set of transliteration rules to convert the Tunisian dialect Arabizi texts to Arabic script. In the second model, transliteration is performed both at word and character levels. In the end, our models got a character error rate of 10.47%.

Understanding what the users say in chatbots: A case study for the Vietnamese language

This paper11This paper is an improved and extended version of Tran and Luong. presents a study on understanding what the users say in chatbot systems: the situation where users input utterances bots would hopefully (1) detect intents and (2) recognize corresponding contexts implied by utterances. This helps bots better understand what users are saying, and act upon a much wider range of actions. To this end, we propose a framework which models the first task as a classification problem and the second one as a two-layer sequence labeling problem. The framework explores deep neural networks to automatically learn useful features at both character and word levels. We apply this framework to building a chatbot in a Vietnamese e-commerce domain to help retail brands better communicate with their customers. Experimental results on four newly-built datasets demonstrate that deep neural networks could be able to outperform strong conventional machine-learning methods. In detecting intents, we achieve the best F-measure of 82.32%. In extracting contexts, the proposed method yields promising F-measures ranging from 78% to 91% depending on specific types of contexts.

Inter-character Orthographic Similarity Effects on the Recognition of Chinese Coordinative Compound Words.

Chinese coordinative compound words are common and unique in inter-character semantic and orthographic relationships. This study explored the inter-character orthographic similarity effects on the recognition of transparent two-morpheme coordinative compound words. Seventy-two native Chinese readers participated in a lexical decision task. The findings demonstrated robust inhibitory inter-character orthographic similarity effects, intra-word character reversal effects, and inter-character semantic similarity effects. These results were compared to those of previous studies on coordinative compound word recognition and on the orthographic similarity phenomenon at both character and word levels. The findings were explained with the multi-level representational model of morphological processing of Chinese compound words (Zhou and Marslen-Wilson in Psychologia 43:47-66, 2000). The model was further extended by adding the activation of morpho-orthographic relationships and the mapping of morphemic orthographic information onto the semantic information of both morphemes and whole words.

Linear Kernel with Weighted Least Square Regression Co-efficient for SVM Based Tamil Writer Identification

Tamil writer identification is the task of identifying writer based on their Tamil handwriting. Our earlier work of this research based on SVM implementation with linear, polynomial and RBF kernel showed that linear kernel attains very low accuracy compared to other two kernels. But the observation shows that linear kernel performs faster than the other kernels and also it shows very less computational complexity. Hence, a modified linear kernel is proposed to enrich the performance of the linear kernel in recognizing the Tamil writer. Weighted least square parameter estimation method is used to estimate the weights for the dot products of the linear kernel. SVM implementation with modified linear kernel is carried out on different text images of handwriting at character, word and paragraph levels. Comparing the performance with linear kernel, the modified kernel with weighted least square parameter reported promising results.

Multi-Perspective Relevance Matching with Hierarchical ConvNets for Social Media Search

Despite substantial interest in applications of neural networks to information retrieval, neural ranking models have mostly been applied to “standard” ad hoc retrieval tasks over web pages and newswire articles. This paper proposes MP-HCNN (Multi-Perspective Hierarchical Convolutional Neural Network), a novel neural ranking model specifically designed for ranking short social media posts. We identify document length, informal language, and heterogeneous relevance signals as features that distinguish documents in our domain, and present a model specifically designed with these characteristics in mind. Our model uses hierarchical convolutional layers to learn latent semantic soft-match relevance signals at the character, word, and phrase levels. A poolingbased similarity measurement layer integrates evidence from multiple types of matches between the query, the social media post, as well as URLs contained in the post. Extensive experiments using Twitter data from the TREC Microblog Tracks 2011–2014 show that our model significantly outperforms prior feature-based as well as existing neural ranking models. To our best knowledge, this paper presents the first substantial work tackling search over social media posts using neural ranking models. Our code and data are publicly available.1

Automatic Diacritics Restoration for Tunisian Dialect

Modern Standard Arabic, as well as Arabic dialect languages, are usually written without diacritics. The absence of these marks constitute a real problem in the automatic processing of these data by NLP tools. Indeed, writing Arabic without diacritics introduces several types of ambiguity. First, a word without diacratics could have many possible meanings depending on their diacritization. Second, undiacritized surface forms of an Arabic word might have as many as 200 readings depending on the complexity of its morphology [12]. In fact, the agglutination property of Arabic might produce a problem that can only be resolved using diacritics. Third, without diacritics a word could have many possible parts of speech (POS) instead of one. This is the case with the words that have the same spelling and POS tag but a different lexical sense, or words that have the same spelling but different POS tags and lexical senses [8]. Finally, there is ambiguity at the grammatical level (syntactic ambiguity). In this article, we propose the first work that investigates the automatic diacritization of Tunisian Dialect texts. We first describe our annotation guidelines and procedure. Then, we propose two major models, namely a statistical machine translation (SMT) and a discriminative model as a sequence classification task based on Conditional Random Fields (CRF). In the second approach, we integrate POS features to influence the generation of diacritics. Diacritics restoration was performed at both the word and the character levels. The results showed high scores of automatic diacritization based on the CRF system (Word Error Rate (WER) 21.44% for CRF and WER 34.6% for SMT).

PENERAPAN LAYANAN BIMBINGAN KELOMPOK DALAM MEMBENTUK KARAKTER SISWA KELAS XI DI SMAN 1 BAITUSSALAM ACEH BESAR

Character is a character (behavior) or character (disposition) inherent in a person. In its formation, many character influenced by environmental factors, namely habits in SMAN 1 Baitussalam Aceh Besar many students who point to bad characters such as lack of respect for the teacher, indifference, ditching and smoking, this the consequences and lack of supervision of student habits. The purpose of this study is to find out whether the application of group guiding services can influence the formation of student character. This study uses quantitative techniques with an experimental approach using One Group Pre-test and Post-test Design. The population of this study is 41 students of social sciences class XI and the sample of this study is 5 students who are selected through the purposive side by categorizing students with low character levels. Data was collected through a Likert scale. Instrument analysis includes validity and reliability analysis. Data analysis using SPSS 20 and t-test. The results of this study is indicate a change in the value of student character scale with an average value of 664 post-test, compared with an average score of 574 pre-test. Based on paired sample test analysis, it is known that there are differences before and after the provision of group guidance services to the character of students in SMAN 1 Baitussalam Aceh Besar, with data t count> t table (13.059> 2.131) or Sig <0.00 so that Ha is accepted and Ho is rejected. This means that group guidance services provided as a treatment for changing student character have a positive effect.

Empower event detection with bi-directional neural language model

Event detection is an essential and challenging task in Information Extraction (IE). Recent advances in neural networks make it possible to build reliable models without complicated feature engineering. However, data scarcity hinders their further performance. Moreover, training data has been underused since majority of labels in datasets are not event triggers and contribute very little to the training process. In this paper, we propose a novel multi-task learning framework to extract more general patterns from raw data and make better use of the training data. Specifically, we present two paradigms to incorporate neural language model into event detection model on both word and character levels: (1) we use the features extracted by language model as an additional input to event detection model. (2) We use a hard parameter sharing approach between language model and event detection model. The extensive experiments demonstrate the benefits of the proposed multi-task learning framework for event detection. Compared to the previous methods, our method does not rely on any additional supervision but still beats the majority of them and achieves a competitive performance on the ACE 2005 benchmark.

ANALYSIS OF CHARACTER VALUES THROUGH ITP AND ATP TECHNOLOGY AS ONE OF THE INNOVATIONS IN LEARNING ELEMENTARY SCHOOL EDUCATION TO OVERCOME DEGRADATION

&lt;em&gt;The development of time because of the existence of globalization led to moral degradation in children. To overcome the moral degradation of the child, it is necessary to have learning innovations in improving the quality of basic education which requires an analysis of character education using methods of observation and questionnaire by entering the results of analysis into ITP software (Development Task Inventory) and ATP (Analysis of Development Task). The existence of ITP and ATP can identify the level of development of aspects related to character education in primary school. Based on preliminary observations, it was found that many teachers did not know the level of children's character development. The article aims to: (1) educators can identify the level of students character (2) educators can design good learning based on the results of each student's character analysis. The conclusion of this study is that educators can develop children's character levels.&lt;/em&gt;

Sense classification of shallow discourse relations with focused RNNs.

Understanding the sense of discourse relations between segments of text is essential to truly comprehend any natural language text. Several automated approaches have been suggested, but all rely on external resources, linguistic feature engineering, and their processing pipelines are built from substantially different models. In this paper, we introduce a novel system for sense classification of shallow discourse relations (FR system) based on focused recurrent neural networks (RNNs). In contrast to existing systems, FR system consists of a single end-to-end trainable model for handling all types and senses of discourse relations, requires no feature engineering or external resources, is language-independent, and can be applied at the word and even character levels. At its core, we present our novel generalization of the focused RNNs layer, the first multi-dimensional RNN-attention mechanism for constructing text/argument embeddings. The filtering/gating RNN enables downstream RNNs to focus on different aspects of the input sequence and project it into several embedding subspaces. These argument embeddings are then used to perform sense classification. FR system has been evaluated using the official datasets and methodology of CoNLL 2016 Shared Task. It does not fall a lot behind state-of-the-art performance on English, the most researched and supported language, but it outperforms existing best systems by 2.5% overall results on the Chinese blind dataset.

Anthracene‐Fused Dibenzo[def,mno]chrysenes with a Helical Structure

AbstractScholl‐type oxidative cyclodehydrogenation reactions were conducted for an anthryl‐substituted dibenzo[def,mno]chrysene (DBC) precursor with the intention to prepare anthracene‐fused DBCs with extended conjugation. The compound ADBC‐1 with one cyclized ring and its two chloro‐ and oxy‐ derivatives (ADBC‐Cl and ADBC‐O) with good stability were successfully isolated and characterized. They showed largely red‐shifted absorption spectra and amphoteric redox behavior. XRD analysis revealed that they can be regarded as large polycyclic aromatic hydrocarbons (PAHs) containing [5]helicene substructures with large twist angles of 67.8°–73.7° owing to steric congestion. Further cyclization to form newer C−C bonds failed, which can be explained by an open‐shell diradical character and high HOMO energy levels of the products according to spin‐unrestricted DFT calculations.

Event Representations for Automated Story Generation with Deep Neural Nets

Automated story generation is the problem of automatically selecting a sequence of events, actions, or words that can be told as a story. We seek to develop a system that can generate stories by learning everything it needs to know from textual story corpora. To date, recurrent neural networks that learn language models at character, word, or sentence levels have had little success generating coherent stories. We explore the question of event representations that provide a mid-level of abstraction between words and sentences in order to retain the semantic information of the original data while minimizing event sparsity. We present a technique for preprocessing textual story data into event sequences. We then present a technique for automated story generation whereby we decompose the problem into the generation of successive events (event2event) and the generation of natural language sentences from events (event2sentence). We give empirical results comparing different event representations and their effects on event successor generation and the translation of events to natural language.

Multi‐dimensional long short‐term memory networks for artificial Arabic text recognition in news video

This study presents a novel approach for Arabic video text recognition based on recurrent neural networks. In fact, embedded texts in videos represent a rich source of information for indexing and automatically annotating multimedia documents. However, video text recognition is a non‐trivial task due to many challenges like the variability of text patterns and the complexity of backgrounds. In the case of Arabic, the presence of diacritic marks, the cursive nature of the script and the non‐uniform intra/inter word distances, may introduce many additional challenges. The proposed system presents a segmentation‐free method that relies specifically on a multi‐dimensional long short‐term memory coupled with a connectionist temporal classification layer. It is shown that using an efficient pre‐processing step and a compact representation of Arabic character models brings robust performance and yields a low‐error rate than other recently published methods. The authors’ system is trained and evaluated using the public AcTiV‐R dataset under different evaluation protocols. The obtained results are very interesting. They also outperform current state‐of‐the‐art approaches on the public dataset ALIF in terms of recognition rates at both character and line levels.

Quinodimethanes Incorporated in Non-Benzenoid Aromatic or Antiaromatic Frameworks.

Three isomers of quinodimethanes (QDMs) adopt different electronic configurations and geometries, generating their own characteristic physical properties. Incorporation of QDMs into non-benzenoid aromatic or antiaromatic frameworks not only planarizes the whole π system optimizing conjugation, but also changes the electronic properties inherent to QDMs, sometimes drastically, due to the topology of the π system, through interaction with the remaining part of the molecules. In non-benzenoid systems, molecular orbital levels and orbital distribution are uneven compared to benzenoid systems, thereby polarizing the ground state and leading to unique behavior in excited states. In antiaromatic systems, open-shell, diradical character, which is inherent to QDMs, may be enhanced due to small HOMO-LUMO energy gap. In this chapter, effects of incorporating QDMs into non-benzenoid aromatic or antiaromatic frameworks are discussed focusing on the open-shell, diradical character with respect to their molecular structures, antiaromaticity, and physical properties related to the open-shell character and molecular orbital levels and materials applications, as well as covering historical works to the current state-of-the-art achievements.

The Influence of Empathy and Morality of Violent Video Game Characters on Gamers' Aggression.

According to the General Aggression Model, situational factors (such as the game characters) and personal factors both affect a gamer’s acquisition of aggressive behavior. Previous studies have found not only that the surface features of game characters, such as appearance and clothing, but also that their inherent characteristics, such as morality and identity, can influence a gamer’s attitude and behavior. Research has also shown that empathy, as a personal factor, can protect gamers from the impact of media violence. However, past research has focused primarily on single factors affecting the player rather than more comprehensive investigations. This study investigates the influence of the game character’s moral features and levels of empathy on the gamer’s aggression. The participants were 120 Chinese university students (61 females and 59 males) with ages ranging from 17 to 27 years. Participants first completed a series of questionnaires: a user experience questionnaire, a video game questionnaire, the Buss-Perry Aggression Questionnaire, and a modified version of the Interpersonal Reactivity Index. All participants then had 5 min of practice playing a violent video game. They were then divided into three groups: a high empathy group, a low empathy group, and a no empathy group. After the practice, participants in the high and low empathy groups read empathy materials relating to the game characters; participants in the no empathy group began formal gameplay. All participants played the game for 20 min. Finally, participants were required to complete the Scale of Hostility Status questionnaire, the Implicit Aggression Test, and the Competitive Reaction Time Test. The results show that empathy and the morality of game characters both influence aggression, but empathy affected aggression differently in the participants playing justified roles (i.e., killing others for a moral reason in the game) compared to those playing unjustified roles (i.e., killing others for immoral reasons in the game). In the high empathy condition, the implicit aggression of justified players was significantly higher than those playing unjustified roles. However, high empathy does not always play a protective role, and its effect is restricted by the features of the game characters.

Learning Simpler Language Models with the Differential State Framework.

Learning useful information across long time lags is a critical and difficult problem for temporal neural models in tasks such as language modeling. Existing architectures that address the issue are often complex and costly to train. The differential state framework (DSF) is a simple and high-performing design that unifies previously introduced gated neural models. DSF models maintain longer-term memory by learning to interpolate between a fast-changing data-driven representation and a slowly changing, implicitly stable state. Within the DSF framework, a new architecture is presented, the delta-RNN. This model requires hardly any more parameters than a classical, simple recurrent network. In language modeling at the word and character levels, the delta-RNN outperforms popular complex architectures, such as the long short-term memory (LSTM) and the gated recurrent unit (GRU), and, when regularized, performs comparably to several state-of-the-art baselines. At the subword level, the delta-RNN's performance is comparable to that of complex gated architectures.

Mediated characters: Multimodal viewpoint construction in comics

AbstractI analyze multimodal viewpoint construction in comics to engage with how modalities function within the medium as a specific discourse context with distinct conventions and material qualities. I show how comics employ established storytelling practices with character, narrator, and narrative viewpoint levels, while building up and interweaving these through strategic uses of the modalities of the medium. I mobilize the cognitive theories of embodiment, domains, mental simulation, and mental space blending as an analytical framework. I examine the asynchronicity of viewpoint elements between modalities and their synthesis into composite character viewpoints in several examples. I show how modalities can be prioritized and their different qualities and functions strategically manipulated for viewpoint construal. These brief examples show the complexity inherent in multimodal communication and interpretation and the usefulness of encouraging the medium-specific and interdisciplinary analyses of cultural works from a cognitive linguistic perspective.

Arabic Online Handwriting Recognition (AOHR)

This article comprehensively surveys Arabic Online Handwriting Recognition (AOHR). We address the challenges posed by online handwriting recognition, including ligatures, dots and diacritic problems, online/offline touching of text, and geometric variations. Then we present a general model of an AOHR system that incorporates the different phases of an AOHR system. We summarize the main AOHR databases and identify their uses and limitations. Preprocessing techniques that are used in AOHR, viz. normalization, smoothing, de-hooking, baseline identification, and delayed stroke processing, are presented with illustrative examples. We discuss different techniques for Arabic online handwriting segmentation at the character and morpheme levels and identify their limitations. Feature extraction techniques that are used in AOHR are discussed and their challenges identified. We address the classification techniques of non-cursive (characters and digits) and cursive Arabic online handwriting and analyze their applications. We discuss different classification techniques, viz. structural approaches, Support Vector Machine (SVM), Fuzzy SVM, Neural Networks, Hidden Markov Model, Genetic algorithms, decision trees, and rule-based systems, and analyze their performance. Post-processing techniques are also discussed. Several tables that summarize the surveyed publications are provided for ease of reference and comparison. We summarize the current limitations and difficulties of AOHR and future directions of research.

Natural scene text detection by multi-scale adaptive color clustering and non-text filtering

In recent years, natural scene text detection gains increasing attention because it plays an important role in many computer related techniques. In this paper, we propose a text detection method consisting of two major steps: connected components (CCs) extraction and non-text filtering. For CCs extraction, a multi-scale adaptive color clustering approach is proposed, which can extract text from images in different color complexities and is robust to contrast variation. For non-text filtering, we combine text covariance descriptor (TCD) with histogram of oriented gradients (HOG) to construct feature vectors and use them to distinguish text from background at character and text line levels. Besides, a new text line generation strategy combining both refined and unrefined CCs is applied, which can retrieve some mis-eliminated characters and generate more integrated text lines. Experiments are conducted on two publicly available datasets, the ICDAR 2013 and the ICDAR 2011 datasets, the obtained F-measures on which are 0.76 and 0.75, respectively. Comparative results with some state-of-the-art text detection algorithms demonstrate that the proposed method achieves competitive performance on text detection.

Adaptive Histogram Analysis for Scene Text Binarization and Recognition

Scene text binarization and recognition is a challenging task due to different appearance of text in clutter background and uneven illumination in natural scene images. In this paper, we present a new method based on adaptive histogram analysis for each sliding window over a word of a text line detected by the text detection method. The histogram analysis works on the basis that intensity values of text pixels in each sliding window have uniform color. The method segments the words based on region growing which studies spacing between words and characters. Then we propose to use existing OCRs such as ABBYY and Tesseract (Google) to recognize the text line at word and character levels to validate the binarization results. The method is compared with well-known global thresholding technique of binarization to show its effectiveness.