Search

Language games represent one of the most fascinating challenges of research in Artificial Intelligence. One of the problems that lies in this set is the development of an automated solution generator for Crossword puzzles. Automatic resolution of crossword puzzles is an open natural language challenge that demands the filling of puzzle grid with candidate answers, while satisfying the grid constraints. Moreover, providing an accurate list of answer candidates has a major impact on the performance of the automatic crossword resolution task. This paper proposes a semantic-based search approach for generating candidate answer lists for definition type clues by taking advantage of the lexical relations encoded in WordNet, the lexical database for English.

Determining semantic similarity between documents is crucial to many tasks such as plagiarism detection, automatic technical survey and semantic search. In this paper, we have mainly focused on detecting the semantic similarity between documents in large documents collection and queries based on an Arabic search engine, we investigated MapReduce as a specific framework for managing distributed processing in dataset pattern and semantic similarity measures of documents. Then we study the state of the art of different approaches for computing the similarity of documents. We propose an approach based on parallel algorithm of semantic similarity measures using MapReduce and WordNet after translation phase to detect the relevant documents in the face of the Arabic query. The numerical results obtained and presented showed the efficiency and the performance of the technique adopted.

Calculating semantic similarity between words is a challenging task of a lot of domains such as Natural language processing (NLP), information retrieval and plagiarism detection. WordNet is a lexical dictionary conceptually organized, where each concept has several characteristics: Synsets and Glosses. Synset represent sets of synonyms of a given word and Glosses are a short description. In this paper, we propose a new approach for calculating semantic similarity between two concepts. The proposed method is based on set theory’s concepts and WordNet properties, by calculating the relatedness between the synsets’ and glosses’s of the two concepts.

Effective geographical query formulation is one of the key difficulties for users in search engines. Our main motivation in this paper is to identify an opportunity for improving geographical search and related technologies. Thus, this work introduces, an approach for geographical query reformulation (GQR), which combines a method of queries’ components separation that uses GeoNames, and a technique for reformulating these components using WordNet and a geographic taxonomy constructed using the Latent Semantic Analysis (LSA) method. In our experiments, we varied the taxonomies construction parameters, like the number of used documents, then we compare the values that are returned in every case. Furthermore, we compared the results of our approach to two methods from the literature, using the Mean Average Precision (MAP) and the precision at 10 documents (P@10). According to our results, reformulating geographical queries using the new proposed approach improves considerably the precision of queries and retrieves relevant documents that were not retrieved using the original queries.

Building of sense-tagged data is a main challenge for supervised techniques that achieved promising results in word sense disambiguation. The manual building of sense-tagged data is a labor and a time-consuming task because each ambiguous word has to be labeled in collected contexts by linguistic experts. Therefore, this paper proposes a knowledge-based method for building the Arabic sense-tagged corpus from Wikipedia. The method starts with mapping Arabic WordNet and Wikipedia to select the Wikipedia article for the corresponding sense in WordNet. In this mapping step, the cross-lingual method is used to measure the similarity between features of a Wikipedia article and a WordNet sense separately. Then, the incoming-links of Wikipedia articles are exploited to extract instances for the sense of each ambiguous word in WordNet. For handling the lack of instances of some articles in Wikipedia, the multiword-based technique is proposed to increase a number of instances for each concept. Experimental results show that the cross-lingual method outperforms monolingual method that is based on Arabic features only. The sense-tagged corpus is created for 50 ambiguous words yielding 148 senses with 30,961 instances.

The dimensionality reduction is a critical problem in the information retrieval process. The higher dimensions directly affect the search performance in terms of Recall and Precision. The dimensionality reduction enabling the search to be semantically based instead of lexically based as the dimensions are defined in terms of the semantic concepts instead of traditional terms or keywords. Latent Semantic Indexing (LSI) is a mathematical extension of the classical Vector Space Model (VSM). LSI is used to discover the latent semantic in the search space by extracting concepts from the original terms in the space. LSI is based on the Singular Value Decomposition (SVD) to reduce the dimension of the term space into a lower dimensional LSI space. In this paper, we propose a methodology for extra optimal LSI dimension reduction via two reduction levels. The first reduction level is based on an ontological conceptualization process. The Universal Wordnet ontology (UWN) is used to develop an ontological based concept space instead of the term space. As a second reduction level, the SVD is applied to the extracted concept space for getting an optimal LSI conceptualization. The experimental results of this research indicate an improvement in the search results in terms of both Precision and Recall as the proposed methodology addresses the Synonymy and Polysemy problems effectively.

Word Sense Disambiguation (WSD) is a task which aims to identify the meaning of a word given its context. This problem has been investigated and analyzed in depth in English. However, work in Arabic has been limited despite the fact that there are half a billion native Arabic speakers. In this work, we present multiple approaches for the problem of WSD in Arabic utilizing recent developments and successes in learning word embeddings with approaches such as GloVe, and Word2vec. The primary shortcoming of word embeddings is the single vector representation of a word’s meaning, although many words are polysemous. Our main contribution in this work is to computationally obtain an embedding for each sense, using an Arabic WordNet (AWN) to overcome the problem of WSD. We also compute word semantic similarity giving thought to multiple Arabic stemming algorithms. Finally, we make available a large pre-processed corpus that is ready to be used for further experiments and a WSD test data based on AWN,1 seeking to fill gaps in Arabic NLP (ANLP) compared to English.

Student profile describes the best way a student prefers to learn. It includes information on student’s characteristics such as background knowledge, learning preference, styles, interest, goals etc. The major challenge that the students face in learning system is that they are unable to retrieve relevant information based on their requirements. One of the methods used to obtain the requirement of the students is to construct an efficient student profile which would reflect the true student needs. The proposed work is to develop an intelligent ontology-based dynamic student profile that provides semantic retrieval using fuzzy concepts. The approach starts with the collection of both static and dynamic data of students. The dynamic dataof students particularly student interest and learning style are obtained by weblog analysis using algorithms such as semantic based representation using WordNet and decision tree classifier algorithm based on Felder-Silverman learning style model (FSLSM). The retrieved data is then used to construct student profile using ontology in which automatic student profile updating is obtained using ontology -based semantic similarity algorithm using WordNet. Finally, semantic retrieval of student information from ontology is achieved by integrating fuzzy concepts using fuzzy linguistic variable and ‘fuzzy IF THEN’ rules. Fuzzy linguistic variable is used to make precise representation on the existing ontology concepts which facilitate more specific classification and semantic retrieval of information. The predictive model of student profile is designed with the implementation of ‘fuzzy IF THEN’ rules using forward chaining reasoning process in the existing ontology model. The inference engine predicts the preference of a new student based on the reasoning process done for specific conditions particularly on student interest and learning style. The experiments were performed using NetBeans IDE, OWL API and Protege 4.2 beta editor. The experiment result shows the successful completion of student profile generation, updating, fuzzy semantic retrieval and prediction through utilization of fuzzy concepts in student profile ontology.

Sentiment analysis is one of the new challenges appeared in automatic language processing with the advent of social networks. Taking advantage of the amount of information is now available, research and industry have sought ways to automatically analyze sentiments and user opinions expressed in social networks. In this paper, we place ourselves in a difficult context, on the sentiments that could thinking of suicide. In particular, we propose to address the lack of terminological resources related to suicide by a method of constructing a vocabulary associated with suicide. We then propose, for a better analysis, to investigate Weka as a tool of data mining based on machine learning algorithms that can extract useful information from Twitter data collected by Twitter4J. Therefore, an algorithm of computing semantic analysis between tweets in training set and tweets in data set based on WordNet is proposed. Experimental results demonstrate that our method based on machine learning algorithms and semantic sentiment analysis can extract predictions of suicidal ideation using Twitter Data. In addition, this work verify the effectiveness of performance in term of accuracy and precision on semantic sentiment analysis that could thinking of suicide.

Hotel booking websites use online ratings and customer feedback to help the customers decision making process but reviews provide a better insight about the hotel but most travellers dont have the time or patience to read all reviews. This study analyzes the hotel reviews and gives information that ratings might overlook. The reviews and metadata are crawled from website and classified into predefined classes as per some of the common aspects. Then Topic modelling technique (LDA) is applied to identify hidden information and aspects, followed by sentiment analysis on classified sentences and summarization. Finally we discuss results and future work, ultimately building towards Hotel Recommender System.

This paper’s focus is to continue the research on Internet psychological warfare analysis, where the authors faced a necessity to propose an accurate algorithm for news clustering that could be able to group news into semantically close sets. A two stage approach to reach that goal is proposed. First a similarity estimation between news messages is performed using semantic similarity metric based on WordNet. Next, the most suitable for given data structure clustering algorithms is selected in order to obtain thematic news clusters and observe their size distribution over time. Experiments were made on news volumes from several news mass media official pages in Facebook.

Most Arabic computational linguistics researchers have focused on Modern Standard Arabic. Linguistic resources and tools for Classical Arabic, especially Al-Hadith Al-Shareef (i.e. the Prophet Muhammad’s saying), remain relatively unexplored, despite its importance as a reference, in addition to its use in the holy Qur’an, used by all Muslims worldwide. Computational linguistics research tools for Al-Hadith Al-Shareef would be useful for both Islamic scholars and learners. Arabic WordNet is a remarkable language resource on its own, allowing a user to determine the relationships among words. It has proven its importance in many of language processing tasks needing an understanding of the meaning of language, including Information Retrieval, Word Sense Disambiguation, Machine Translation, Question Answering, Text Classification, and Text Summarization. In this paper, we propose an approach for developing a WordNet linguistic resource for Al-Hadith Al-Shareef that serves its purposes for various Arabic natural language processing tasks. In particular, we establish semantic connections between words in order to achieve a good understanding of the meanings of the Al-Hadith words. Our approach employs Classical Arabic dictionaries and Al-Hadith ontology. Al-Hadith WordNet has demonstrated its capability in a text classification task that we developed for evaluation proposes. The classifier has been applied on around 8500 synsets that include 6126 nominal, 1990 verbal, 310 adjectival, and 71 adverbial expressions.

The value of big data comes from its variety where data is collected from various sources. One of the key big data challenges is identifying which data objects are relevant or refer to the same logical entity across various data sources. This challenge is traditionally known as schema matching. Due to big data velocity traditional approaches to data matching can no longer be used. In this paper we present an approach for inferring data objects correlation. We present our algorithm that relies on the objects meta-data and it consults the Word Net thesaurus

Document annotation and search are two important factors that need to be considered in data sharing platforms. A large unstructured text document contains substantial amount of structured attribute information. Important information is very difficult to find in these documents. Current ad-hoc or predefined annotation of the shared data causes inadequate search, retrieval and analysis capabilities. In this paper we propose a new approach that supports the generation of the structured annotation in the form of attribute name and attribute value pairs from unstructured document. A new data sharing platform DSPAA (Data Sharing Platform with Automated Annotation) is proposed, where the document annotation occurs when the author uploads a document and it is based on a probabilistic framework that considers the attributes in the document content and the query collection. The new system also performs semantic annotation of document using WordNet database. When a user submits a search query, then the system will search for documents in the annotation database and rank the selected documents by using Vector Space model. From experiment results it is clear that the system generates superior results at a rate faster than traditional document retrieval strategies.

In this paper, a novel approach using Description Logic (DL) based inference rules, for ontology matching is presented. Alignment concerns ontology concepts, with the application of similarity measures to perform concepts and instances relationship alignments. Moreover, external knowledge, in the form of WordNet dictionary is then used to solve usual matching problems encountered with synonyms, polesemy, homonyms, etc. Illustrative examples are then presented to support the developed approach.

The purpose of this paper is to present an approach to create semi-automatically ontology from Arabic texts. The whole process is supervised by a linguistic expert. Our involvement in this project focused on a lexical ontology, taking as model the WordNet ontology, and as input source, the “Arabic verbs” of a contemporary monolingual dictionary ( ) /mζjm Alγny/ in the form a lexical database. The verb, pivot of a sentence, is our goal in creating concepts, by adopting the synset as our meaning representation model. The Markov clustering algorithm of a graph, generated by the defining verbs, obtained from the transitive closure, allowed us to detect similar verbs and to identify as well, for a given verbal entry, all of its synonyms. A tool has been implemented, and experiments have been carried out to evaluate and show efficiency of the proposed approach.

One of the greatest challenges in speech technology is estimating the speaker's emotion. Most of the existing approaches concentrate either on audio or text features. In this work, we propose a novel approach for emotion classification of audio conversation based on both speech and text. The novelty in this approach is in the choice of features and the generation of a single feature vector for classification. Our main intention is to increase the accuracy of emotion classification of speech by considering both audio and text features. In this work we use standard methods such as Natural Language Processing, Support Vector Machines, WordNet Affect and SentiWordNet. The dataset for this work have been taken from Semval -2007 and eNTERFACE’05 EMOTION Database.

A novel approach based on supervised classification has been proposed to classify a given collection of XML documents based on rule based classifier by semantically enriched structure and content features. The proposed methodology overcomes the drawbacks of the existing technologies by accomplishing the classification by utilizing not only the structure and content features but also context. It applies ontological information into structural and content based features from the XML documents and transforms it into transaction formats onto which FP-growth algorithm is executed to generate association rules. An associative classifier is constructed by eliminating irrelevant rules from the generated association rule.

Rule based systems have achieved success in applications such as information retrieval and Natural Language Processing. However, due to the rigidity of pattern matching, these systems typically require a large number of rules to adequately cover the variations of expression in unstructured text. Consequently, knowledge engineering for a new domain and knowledge maintenance for a fielded system are labor intensive and expensive. In this paper, we present our research on enhancing a rule-based event coding system by relaxing the rigidity of pattern matching with a technique that formulates and matches patterns of the semantics of words instead of literal words. Our technique pairs literal words with semantic vectors that accumulate word meaning from the context of use of the word found in dictionaries, ontologies, and domain corpora. Our method improves the speed, accuracy, and coverage of the event coding algorithm without additional knowledge engineering effort. Operating on semantics instead of syntax, the improved system eases the workload of human analysts who screen input text for critical events. Our algorithms are based on high-dimensional distributed representations, and their effectiveness and versatility derive from the unintuitive properties of such representations---from the mathematical properties of high-dimensional spaces. Our current implementation encodes words, phrases, and rule patterns as semantic vectors using WordNet, We have started experimental evaluation using a large newswire dataset.

The challenges of machine semantic understanding have not yet been satisfactorily solved by automated methods. In our approach, the semantics and syntax of words, phrases and documents are represented by deep semantic vectors that capture both the structure and semantic meaning of the language. Our experiment reproduces the experiment done by Patwardhan and Pedersen 2006, but uses random index vectors for the words, glosses and tweets. Our model first determines random index vectors from glosses and definitions for words from WordNet. From these foundational semantic vectors, random index vectors that represent phrases, sentences or tweets are determined. Our set of algorithms relies on high-dimensional distributed representations, and their effectiveness and versatility derive from the unintuitive properties of such representations: from the mathematical properties of high-dimensional spaces. High-dimensional vector representations have been used successfully in modeling human cognition, such as memory and learning. Our semantic vectors are high-dimensional and capture the meaning of a language expression, such as a word, phrase, query, news article, story or a message. A key benefit of our method is that the dimensionality of the vectors remains constant as we add data; this also allows good generalization to rarely seen words, which “borrow strength” from their more frequent neighbors.

Web page retrieval plays a major role nowadays. The proposed system helps in the development of optimized web page recommendation. The proposed system is a novel method to provide the web pages effectively and accurately by using genetic algorithm. First the entered query is pre-processed and similar word set is generated. It uses word net tool for semantic keyword set generation that is the various fields where that particular word is used. This tool uses the ontology concept for the keyword generation. The user has to click a word from these keyset or directly click the pre-processed word. The selected keyword is given as input to server and some webpages will be retrieved. The web contents are extracted from these webpages. Then the weightage of that particular word is calculated in all the documents/web contents retrieved. Weight Enhanced genetic algorithm which involves three main process that is selection, crossover and mutation is applied to the weights calculated. Finally, the webpage with highest weightage will be displayed first and then consequently remaining webpages will be displayed in the same manner.

The need of humans to socialize and share information has led to a constantly growing Web, which has become a support for social media. Every day, worldwide users are pushing multimedia data towards their family, friends and the world at large. This is the reason why, web search has also become the main method for people to fulfill their information needs. The common modality used for image search on the web is based on text, the assumption being that the tags and the textual descriptions associated with photos are powerful ways to describe and retrieve images. The results are usually obtained by simply matching the terms of the query to an index of terms associated to the images in a corpus. The efficiency of this technique depends strongly on the tags associated to pictures, as well as their accuracy. Trying to search for images with bridges in this kind of systems, the result set will only contain images explicitly annotated with this term, but will fail to include images with Pont Neuf or Ponte Vecchio, if their tags do not contain the noun bridge. In this paper, we present a system designed to perform diversification in an image retrieval system, using semantic resources like YAGO, Wikipedia and WordNet.

Semantically enriched multimedia information is crucial for equipping the kind of multimedia search potentials that professional searchers need. But the semantic interpretation of multimedia is obsolete without some mechanism for understanding semantic content that is not explicitly available. Manual annotation is the only source to overwhelming this, which is not only time consuming and costly but also lacks semantic enrichment in terms of concept diversity and concept enrichability. In this paper, we present semantically enhanced information extraction model that calculate the semantic intensity (SI) of each object in the image and then enhance the tagged concept with the assistance of lexical and conceptual knowledgebases .i.e. WordNet and ConceptNet. Noises, redundant and unusual words are then filtered out by means of various techniques like semantic similarity, stopwords and words unification. The experiment has been carried out on the LabelMe datasets. Results demonstrate the substantial improvement in terms of concept diversity, concept enrichment and retrieval performance.

Patent search is a complex task and involves a great level of expertise. Through this research we have tried to find similar patents by expanding the user query semantically. The main purpose of this research is to investigate how the patent retrieval system can be improved by using words which have same expression i.e. semantically similar. WorldNet and Wikipedia are used as an external source for expanding the query. Result shows that expanded query yields better results compared to conventional approaches of patent search.