Assessing Student Paraphrases Using Lexical Semantics and Word Weighting

Word Sense Disambiguation (WSD) is the process of resolving the meaning of a word unambiguously in a given natural language context. Within the scope of this thesis, it is the process of marking text with explicit sense labels. What constitutes a sense is a subject of great debate. An appealing perspective, aims to define senses in terms of their multilingual correspondences, an idea explored by several researchers, Dyvik (1998), Ide (1999), Resnik & Yarowsky (1999), and Chugur, Gonzalo & Verdejo (2002) but to date it has not been given any practical demonstration. This thesis is an empirical validation of these ideas of characterizing word meaning using cross-linguistic correspondences. The idea is that word meaning or word sense is quantifiable as much as it is uniquely translated in some language or set of languages. Consequently, we address the problem of WSD from a multilingual perspective; we expand the notion of context to encompass multilingual evidence. We devise a new approach to resolve word sense ambiguity in natural language, using a source of information that was never exploited on a large scale for WSD before. The core of the work presented builds on exploiting word correspondences across languages for sense distinction. In essence, it is a practical and functional implementation of a basic idea common to research interest in defining word meanings in cross-linguistic terms. We devise an algorithm, SALAAM for Sense Assignment Leveraging Alignment And Multilinguality, that empirically investigates the feasibility and the validity of utilizing translations for WSD. SALAAM is an unsupervised approach for word sense tagging of large amounts of text given a parallel corpus—texts in translation—and a sense inventory for one of the languages in the corpus. Using SALAAM, we obtain large amounts of sense annotated data in both languages of the parallel corpus, simultaneously. The quality of the tagging is rigorously evaluated for both languages of the corpora. The automatic unsupervised tagged data produced by SALAAM is further utilized to bootstrap a supervised learning WSD system, in essence, combining supervised and unsupervised approaches in an intelligent way to alleviate the resources acquisition bottleneck for supervised methods. Essentially, SALAAM is extended as an unsupervised approach for WSD within a learning framework; in many of the cases of the words disambiguated, SALAAM coupled with the machine learning system rivals the performance of a canonical supervised WSD system that relies on human tagged data for training. Realizing the fundamental role of similarity for SALAAM, we investigate different dimensions of semantic similarity as it applies to verbs since they are relatively more complex than nouns, which are the focus of the previous evaluations. We design a human judgment experiment to obtain human ratings on verbs' semantic similarity. The obtained human ratings are cast as a reference point for comparing different automated similarity measures that crucially rely on various sources of information. Finally, a cognitively salient model integrating human judgments in SALAAM is proposed as a means of improving its performance on sense disambiguation for verbs in particular and other word types in general.

Similarity calculation between Business Process Models (BPM) has an important role in the process of managing BPM repository. One of its uses is to facilitate the searching process of a model in the repository. Similarity calculation between business processes is closely related with semantic string similarity. Semantic string similarity is usually performed by utilizing a lexical database, such as WordNet, to find the semantic meaning of words. The problem in WordNet is that this lexical database contains terms wich have more than one meaning or polysemous. Selecting the wrong meaning will decrease the accuracy of similarity calculation process. In this study, we will try to improve the accuracy of similarity calculation of business processes using Word Sense Disambiguation (WSD). The main purpose is to eliminate the ambiguity of polysemous words before calculating the similarity value. WSD is performed by unsupervised methods based on the value of graph connectivity. Then, we used a lexical database that is focused in the business and industry field. The results from this study is able to achieve higher accuracy of the sense selection process for terms especially terms that are related to business and industrial domains. It will also increase the accuracy of similarity value calculation between the business process models .

This paper presents a method of word sense disambiguation that assigns a target word the sense that is most related to the senses of its neighbor words. We explore the use of measures of relatedness between word senses based on a novel hybrid approach. First, we investigate how to “literally” and “regularly” express a “concept”. We apply set algebra to Wordnet's synsets cooperating with Wordnet's word ontology. In this way we establish regular rules for constructing various representations (lexical notations) of a concept using Boolean operators and word forms in various synset(s) defined in Wordnet. Then we establish a formal mechanism for quantifying and estimating the semantic relatedness between concepts – we facilitate “concept distribution statistics” to determine the degree of semantic relatedness between two lexically expressed concepts. Human languages have words that can mean different things in different contexts, such words with multiple meanings are potentially “ambiguous”. The process of “deciding which of several meanings of a term is intended in a given context” is known as “Word Sense Disambiguation (WSD)”. The proposed method is not supervised, and does not require any manually created sense‐tagged training examples. The experimental results showed good performance on Semcor, a subset of the Brown corpus. We observe that measures of semantic relatedness are useful sources of information for word sense disambiguation.

This paper presents a method of word sense disambiguation that assigns a target word the sense that is most related to the senses of its neighbor words. We explore the use of measures of relatedness between word senses based on a novel hybrid approach. First, we investigate how to literally and regularly express a concept. We apply set algebra to Wordnet’s synsets cooperating with Wordnet’s word ontology. In this way we establish regular rules for constructing various representations (lexical notations) of a using Boolean operators and various word forms in synset(s). Then we construct a formal mechanism for quantifying and estimating the semantic relatedness between concepts - we facilitate concept distribution statistics to determine the degree of semantic relatedness between two lexically expressed concepts. Then we applied the measure of semantic relatedness to the WSD task. The experimental results showed good performance on Semcor, a subset of Brown corpus. We observe that measures of semantic relatedness are useful sources of information for word sense disambiguation.

The research presented in this paper builds on previous work that lead to the definition of a family of semantic relatedness algorithms that compute a proximity given as input a pair of concept labels. The algorithms depends on a semantic graph, provided as RDF data, and on a particular set of weights assigned to the properties of RDF statements (types of arcs in the RDF graph). The current research objective is to automatically tune the weights for a given graph in order to increase the proximity quality. The quality of a semantic relatedness method is usually measured against a benchmark data set. The results produced by the method are compared with those on the benchmark using the Spearman's rank coefficient. This methodology works the other way round and uses this coefficient to tune the proximity weights. The tuning process is controlled by a genetic algorithm using the Spearman's rank coefficient as the fitness function. The genetic algorithm has its own set of parameters which also need to be tuned. Bootstrapping is based on a statistical method for generating samples that is used in this methodology to enable a large number of repetitions of the genetic algorithm, exploring the results of alternative parameter settings. This approach raises several technical challenges due to its computational complexity. This paper provides details on the techniques used to speedup this process. The proposed approach was validated with the WordNet 2.0 and the WordSim-353 data set. Several ranges of parameters values were tested and the obtained results are better than the state of the art methods for computing semantic relatedness using the WordNet 2.0, with the advantage of not requiring any domain knowledge of the ontological graph.

All human languages have words that can mean different things in different contexts, such words with multiple meanings are potentially “ambiguous”. The process of “deciding which of several meanings of a term is intended in a given context” is known as “word sense disambiguation (WSD)”. This paper presents a method of WSD that assigns a target word the sense that is most related to the senses of its neighbor words. We explore the use of measures of relatedness between word senses based on a novel hybrid approach. First, we investigate how to “literally” and “regularly” express a “concept”. We apply set algebra to WordNet’s synsets cooperating with WordNet’s word ontology. In this way we establish regular rules for constructing various representations (lexical notations) of a concept using Boolean operators and word forms in various synset(s) defined in WordNet. Then we establish a formal mechanism for quantifying and estimating the semantic relatedness between concepts—we facilitate “concept distribution statistics” to determine the degree of semantic relatedness between two lexically expressed concepts. The experimental results showed good performance on Semcor, a subset of Brown corpus. We observe that measures of semantic relatedness are useful sources of information for WSD.

To save time, healthcare providers frequently use abbreviations while authoring clinical documents. Nevertheless, abbreviations that authors deem unambiguous often confuse other readers, including clinicians, patients, and natural language processing (NLP) systems. Most current clinical NLP systems "post-process" notes long after clinicians enter them into electronic health record systems (EHRs). Such post-processing cannot guarantee 100% accuracy in abbreviation identification and disambiguation, since multiple alternative interpretations exist. Authors describe a prototype system for real-time Clinical Abbreviation Recognition and Disambiguation (rCARD) - i.e., a system that interacts with authors during note generation to verify correct abbreviation senses. The rCARD system design anticipates future integration with web-based clinical documentation systems to improve quality of healthcare records. When clinicians enter documents, rCARD will automatically recognize each abbreviation. For abbreviations with multiple possible senses, rCARD will show a ranked list of possible meanings with the best predicted sense at the top. The prototype application embodies three word sense disambiguation (WSD) methods to predict the correct senses of abbreviations. We then conducted three experments to evaluate rCARD, including 1) a performance evaluation of different WSD methods; 2) a time evaluation of real-time WSD methods; and 3) a user study of typing clinical sentences with abbreviations using rCARD. Using 4,721 sentences containing 25 commonly observed, highly ambiguous clinical abbreviations, our evaluation showed that the best profile-based method implemented in rCARD achieved a reasonable WSD accuracy of 88.8% (comparable to SVM - 89.5%) and the cost of time for the different WSD methods are also acceptable (ranging from 0.630 to 1.649 milliseconds within the same network). The preliminary user study also showed that the extra time costs by rCARD were about 5% of total document entry time and users did not feel a significant delay when using rCARD for clinical document entry. The study indicates that it is feasible to integrate a real-time, NLP-enabled abbreviation recognition and disambiguation module with clinical documentation systems.

Non-Learning Semantic Analysis for Context Discovery and Sentiment Estimation: Transportation Application by Himanshu Verma 〈Dr. Pushkin Kachroo〉, Examination Committee Chair Professor of Electrical and Computer Engineering University of Nevada, Las Vegas With enormous amount of linguistic data present on web, text analysis has become one of the major fields of interest today. This field includes sentiment analysis, information retrieval, text document classification, knowledge based modeling, content similarity measure, data clustering, words prediction/correction, decision making etc. Managing and processing such data has vital importance. The field being quite broad, our focus is mainly on transportation related social media(Twitter) data extraction, text categorization/classification which can be further sub-divided into concept discovery, word sense disambiguation and sentiment analysis to analyze performance of existing transportation system worldwide. Concept discovery is the method of extracting the actual concept/context in which the text is about. This also allows us to filter irrelevant data. Word sense disambiguation is to find the correct sense in which a word is being used in a sentence. It is the basic necessity for concept discovery. iii A lot of research has been done in this field with major improvements. However, when it comes to short texts, the field still seems in nascent stage. Moreover, most of the methods today require huge amount training corpus(database). Arranging such corpus is a cumbersome task and requires a lot of human effort. The other problem with the existing methods are that they require a set of defined concepts from which a concept is chosen and labeled to text. We will consider the case of finding a general context. In this work a novel approach has been proposed for word sense disambiguation which in turn allows us to find general context. For this purpose, I have used the existing knowledge based semantic dictionary called WordNet. This methodology helps in avoiding the use of huge corpus and works for general context recognition. Our focus is on short-texts(Tweets) but the concept is easily applicable to text documents as well. Sentiment measuring technique was applied on extracted data and the scores were mapped to google maps based on the location information present in the tweets. This clearly points out the locations where people are more frustrated with the existing transportation system and need immediate attention for improvements.

QuantNet is an integrated web-based environment consisting of different types of statistics-related documents and program codes. Its goal is creating reproducibility and offering a platform for sharing validated knowledge native to the social web. To increase the information retrieval (IR) efficiency there is a need for incorporating semantic information. Three text mining models will be examined: vector space model (VSM), generalized VSM (GVSM) and latent semantic analysis (LSA). The LSA has been successfully used for IR purposes as a technique for capturing semantic relations between terms and inserting them into the similarity measure between documents. Our results show that different model configurations allow adapted similarity-based document clustering and knowledge discovery. In particular, different LSA configurations together with hierarchical clustering reveal good results under M3 evaluation. QuantNet and the corresponding Data-Driven Documents (D3) based visualization can be found and applied under quantlet. The driving technology behind it is Q3-D3-LSA, which is the combination of “GitHub API based QuantNet Mining infrastructure in R”, LSA and D3 implementation.

LWCR: multi-Layered Wikipedia representation for Computing word Relatedness

Semantic relatedness and disambiguation are fundamental problems for linking text documents to the Web of Data. There are many approaches dealing with both problems but most of them rely on word or concept distribution over Wikipedia. They are therefore not applicable to concepts that do not have a rich textual description. In this paper, we show that semantic relatedness can also be accurately computed by analysing only the graph structure of the knowledge base. In addition, we propose a joint approach to entity and word-sense disambiguation that makes use of graph-based relatedness. As opposed to the majority of state-of-the-art systems that target mainly named entities, we use our approach to disambiguate both entities and common nouns. In our experiments, we first validate our relatedness measure on multiple knowledge bases and ground truth datasets and show that it performs better than related state-of-the-art graph based measures. Afterwards, we evaluate the disambiguation algorithm and show that it also achieves superior disambiguation accuracy with respect to alternative state-of-the-art graph-based algorithms.

Recently, Wikipedia has become a very important resource for computing semantic relatedness (SR) between entities. Several approaches have already been proposed to compute SR based on Wikipedia. Most of the existing approaches use certain kinds of information in Wikipedia (e.g. links, categories, and texts) and compute the SR by empirically designed measures. We have observed that these approaches produce very different results for the same entity pair in some cases. Therefore, how to select appropriate features and measures to best approximate the human judgment on SR becomes a challenging problem. In this paper, we propose a supervised learning approach for computing SR between entities based on Wikipedia. Given two entities, our approach first maps entities to articles in Wikipedia; then different kinds of features of the mapped articles are extracted from Wikipedia, which are then combined with different relatedness measures to produce nine raw SR values of the entity pair. A supervised learning algorithm is proposed to learn the optimal weights of different raw SR values. The final SR is computed as the weighted average of raw SRs. Experiments on benchmark datasets show that our approach outperforms baseline methods.

Latent semantic analysis (LSA) has been used in several intelligent tutoring systems(ITS's) for assessing students' learning by evaluating their answers to questions in the tutoring domain. It is based on word-document co-occurrence statistics in the training corpus and a dimensionality reduction technique. However, it doesn't consider the word-order or syntactic information, which can improve the knowledge representation and therefore lead to better performance of an ITS. We present here an approach called Syntactically Enhanced LSA (SELSA) which generalizes LSA by considering a word along with its syntactic neighborhood given by the part-of-speech tag of its preceding word, as a unit of knowledge representation. The experimental results on Auto-Tutor task to evaluate students' answers to basic computer science questions by SELSA and its comparison with LSA are presented in terms of several cognitive measures. SELSA is able to correctly evaluate a few more answers than LSA but is having less correlation with human evaluators than LSA has. It also provides better discrimination of syntactic-semantic knowledge representation than LSA.

In recent days, semantic web presents a key role in intelligent retrieval of information system that resolves vocabulary mismatch problem by query expansion process. However, achieving the scalable information retrieval (IR) in semantic web is a challenging issue in a large dataset. The semantic IR problem is addressed by an ontological-based semantic similarity measurement using natural language processing. The two novel algorithms namely syntactic correlation coefficient (SCC) and mapping-based K-nearest neighbour (M-KNN) for semantic similarity measurement is proposed which improves the accuracy of relevant result. The ontological constructs with word sense disambiguation (WSD) algorithm for document repository improves the conceptual relationships, reduces the ambiguities in ontology and improves scalability by intensely analysing the semantic relationship as well as dynamically reconstructing the ontology when numbers of documents are updated. Ranking is done with latent semantic analysis (LSA) after semantic similarity analysis, which improves the retrieved result and reduces the complexity in relevancy. The performance of the system is analysed with respect to different metrics such as processing time, F-measure (0.97), time complexity, precision (0.95), recall (0.98) and space complexity.

With the development of search engines (e.g. Google, Bing, Yahoo, etc.), people is ambitiously expecting higher quality and improvements of current technologies. Bringing human intelligence features to these tools, like the ability to find implicit information through semantics, is one of the must prominent research lines in Computer Science. Information semantics is a very wide concept, as wide as the human capability to interpret, in particular, the analysis of geographical semantics gives the possibility to associate information with a place. It is estimated that more than 70\% of all information in the world has some kind of geographic features \cite{Jones04}. In 2012, Ed Parsons, a GeoSpatial Technologist from Google, reported that between 30\% and 40\% of the user queries at Google search engine contain geographic references \cite{Parsons12}. This thesis addresses the field of geographic information extraction and retrieval in unstructured texts. This process includes the identification of spatial features in textual documents, the data indexing, the manipulation of the relevance of the identified geographic entities and the multi-criteria retrieval according to the thematic and geographic information. The main contributions of this work include a custom geographic knowledge base, built from the combination of GeoNames and WordNet; a Natural Language Processing and knowledge based heuristics for Toponym Recognition and Toponym Disambiguation; and a geographic relevance weighting model that supports non-spatial indexing and simple ranking combination approaches. The validity of each one of these components is supported by practical experiments that show their effectiveness in different scenarios and their alignment with state of the art solutions. In addition, it also constitutes a main contribution of this work GEIR, a general purpose GIR framework that includes the implementations of the above described components and brings the possibility of implementing new ones and test their performance within an end to end GIR system.