Approaches for Word Sense Disambiguation - A Survey

2018

In natural language processing (NLP), word sense disambiguation (WSD) is an automatic process carried out by a machine to sense the appropriate meaning of a word in a particular context or in a discourse. Natural language is ambiguous, so that many words may be interpreted in multiple methods depending on the context wherein they occur. The computational identification of which means for words in context is known as word sense disambiguation (WSD). In this paper, we will discuss the ambiguity of the words in the languages and the essential measures to deal with the ambiguous words.

There is a renewed interest in word sense disambiguation (WSD) as it contributes to various applications in natural language processing. Applications for which WSD is potentially an issue are: Machine Translation, Information Retrieval (IR), QA systems, Dialogue systems,etc. In this paper we survey vector-based methods for WSD in machine learning approache.

International Journal of Computer Applications, 2015

This paper presents various techniques used in the area of Word Sense Disambiguation (WSD). There are a number of techniques such as: Knowledge based approaches, which use the knowledge encoded in Lexical resources; Supervised Machine Leaning methods in which the classifier is made to learn from previously semantically annotated corpus; Unsupervised approaches that form cluster occurrences of words. Then there are also semi supervised approaches which use semi annotated corpus as reference data along with unlabeled data.

2007

Word sense disambiguation (WSD) is the task of selecting the appropriate senses of a word in a given context. It is essence of communication in a natural language. It is motivated by its use in many crucial applications such as Information retrieval, Information extraction, Machine Translation, Partof-Speech tagging, etc. Various issues like scalability, ambiguity, diversity (of languages) and evaluation pose challenges to WSD solutions. The aim of this project is to develop a WSD technique which can handle all these issues with better accuracy and performance. This report presents our preliminary work towards solving the problem.

Word Sense Disambiguation is the process of determining which sense of a word is used in a given context. Word Sense Disambiguation is a classi cation problem. Given an instance of a word and the context in which it occurs , the aim is to determine the sense of that occurrence of the word . In this paper a comparison of three recent word sense disambiguation tech- niques is presented. The Word Sense Disambiguation methods surveyed are as fol- lows. Word Sense Disambiguation with automatically acquired knowledge, Word Sense Disambiguation using enriched semantic relation network and a Word Sense Disambiguation based on Lexical Chaining. These approaches are evaluated with standard corporas and their results are compared. These approaches are extensively used in Natural Language Processing

The words that are often being correspond to two or more meanings rather than to a single meaning results in semantically-ambiguous words. Measuring the similarity between words, sentences , paragraphs is an important part in information retrieval and word sense disambiguation tasks. One of the biggest challenges in Natural Language Processing is for the system to encompass in what sense a specific word is being used .This paper describes the analysis of text in order to a certain first the similarity in case that exists. Second the effort has been made to resolve the ambiguity in the text. The paper presents the comparison of machine learning approaches in the text similarity analysis. The Naive bayes approach was observed to outperform other approaches including SVM , Max Entropy , Tree , Random Forest and Bagging. Keywords: Text Similarity, Word Sense Disambiguation, Approaches, SENSEVAL, Supervised machine learning algorithms. I. INTRODUCTION In the present world people are mainly depended on the web for searching any kind of content. Search engines have done remarkable job of information retrieval. However, but still the goal of retrieving relevant information is a far cry. When the person is searching information on web he /she does not bother about the ambiguity of a word that whether the content they are retrieving is relevant to them or not. It gets difficult for the user to get relevant information in any language when the word or phrases have more than one interpretation. One step towards realizing this goal is the detection of similarity of texts i.e. determining how close is the meaning of two given texts are. The idea is based on text similarity [1] detection which plays an important role in text related search in tasks such as information retrieval, word sense disambiguation (WSD) , machine translation , Information Extraction and Speech Recognition and others. For example , the phrase " The second hand of the clock is not working " , the word second means a basic unit of time , while in phrase " Ram came second in the class " , the word " second " refers to the position in series .The problem can be reduced up to an extent by the concept of disambiguation of a word. When a word has multiple meaning then it is probably considered an ambiguity. Hence, Word sense disambiguation (WSD) is termed as an open problem of natural language processing with a process of identifying a correct sense of a word in a given context. WSD plays important role in improving the quality of information so as to comprehend in what sense a specific word is being used. WSD was first formulated as a distinct ciphering task during early days of machine translation in late 1940s, making one of the oldest problem of computational semantics. The problem was continued as a challenging task until there was a availability of resources. In 1980 there was prodigious development in the area of WSD research when a large scale lexical resources and corpora came into existence. In 1990s , NLP provided three major developments for WSD :online dictionary WordNet which is organised as a word senses called synsets and used as an online sense inventory ,statistical methodologies which are used as sense classification problems and SENSEVAL which was proposed in 1997 by Resnik and Yarowsky. Further other SENSEVAL evaluation exercises have also been introduced so that researchers can share and upgrade their views in this research area. II. LITERATURE REVIEW When the work started on handling of the different languages with automatic means, the problem of ambiguity drew the interest of the researchers at the same time. Work on ambiguity in sense annotation has often focused on techniques to reduce ambiguity in sense inventory .Therefore, we can say that the WSD task is one of the oldest tasks for solving lexical ambiguity. Many of the researchers[2]Mukti Desai and Mrs. Kiran Bhowmick (2013) have surveyed on solving the ambiguity by applying different approaches and techniques of WSD. [3] A. R. Rezapour et al. (2011)have used a K-Nearest Neighbor algorithm of supervised learning method for WSD. The author have done feature extraction which includes the set of words that have occurred frequently in the text and the set of words surrounding the ambiguous word, so as to improve the classification accuracy. [4]Arti Mishra and Meenakshi Pathak (2014) have analyzed the web queries in English language to study the effect on the performance of various

2006

In the Natural Language Processing (NLP) community, Word Sense Disambiguation (WSD) has been described as the task which selects the appropriate meaning (sense) to a given word in a text or discourse where this meaning is distinguishable from other senses potentially attributable to that word. These senses could be seen as the target labels of a classification problem. That is, Machine Learning (ML) seems to be a possible way to tackle this problem.

Abstract Word Sense Disambiguation (WSD) is one of the most important open problems in Natural Language Processing. One of the most successful current lines of research in WSD is the corpus-based approach, in which machine learning algorithms are applied to learn statistical models or classifiers from corpora. When a machine learning approach learns from previously semantically annotated corpora it is said to be supervised, whereas when it does not use sense tagged data during training it is called unsupervised.

Journal of Information System and Technology Management, 2021

Background: Word Sense Disambiguation (WSD) is known to have a detrimental effect on the precision of information retrieval systems, where WSD is the ability to identify the meanings of words in context. There is a challenge in inference-correct-sensing on ambiguous words. Through many years of research, there have been various solutions to WSD that have been proposed; they have been divided into supervised and knowledge-based unsupervised. Objective: The first objective of this study was to explore the state-of-art of the WSD method with a hybrid method using ontology concepts. Then, with the findings, we may understand which tools are available to build WSD components. The second objective was to determine which method would be the best in giving good performance results of WSD, by analysing how the methods were used to answer specific WSD questions, their production, and how their performance was analysed. Methods: A review of the literature was conducted relating to the performa...

Proceedings of the 37th annual meeting …, 1999

Selecting the most appropriate sense for an ambiguous word in a sentence is a central problem in Natural Language Processing. In this paper, we present a method that attempts to disambiguate all the nouns, verbs, adverbs and adjectives in a text, using the senses provided in WordNet. The senses are ranked using two sources of information: (1) the Internet for gathering statistics for word-word cooccurrences and (2)WordNet for measuring the semantic density for a pair of words. We report an average accuracy of 80% for the first ranked sense, and 91% for the first two ranked senses. Extensions of this method for larger windows of more than two words are considered.

2012 12th International Conference on Hybrid Intelligent Systems (HIS), 2012

Natural language processing applications invariably perform word sense disambiguation as one of its processing steps. The accuracy of sense disambiguation depends upon an efficient algorithm as well as a reliable knowledge-base in the form of annotated corpus and/or dictionaries in machine readable form. Algorithms working on corpus for sense disambiguation are generally employed as supervised machine learning systems. But such systems need ample training on the corpus before being applied on the actual data set. This paper discusses an unsupervised approach of a graph-based technique that solely works on a machine-readable dictionary as the knowledge source. This approach can improve the bottleneck problem that persists in corpus-based word sense disambiguation. The method described here attempts to make the algorithm more intelligent by considering various WordNet semantic relations and auto-filtration of content words before graph generation.

Gi Jahrestagung, 2006

Naturall anguagep rocessing( NLP) hasb ecome them osts ignificant obstaclethathas been restrictingthe applications viathe web. Today, very little of thec ontento nt he webc an be understood by them achines, althoughv ast amount of electronici nformationh as been kept on them.W ords ense disambiguation (WSD)i sa ni mportant intermediate step in many language processing applications.I ti sb asically am apping functionf romt he contextt ot he seto f senses. This functionhas many parameters that aredifficult to explore. Thefactors effectingt he successo fW SD systemsa re generallyv erys ensitivet ot hese parameters. Thei ssues in WSDa re examined in thec ontext of TurkishW SD application. Ambiguous wordsand theirsense classifications have been established andb yp roviding manually senset aggedc orporaa nd examiningW SD problem from various perspectives, an important contributionh as been achievedf or the researches in this domain.

2018

Word Sense Disambiguation has been a research area since the evolution of Natural Language processing. Most of the languages have ambiguous words. Resolution to these ambiguous words is a very importance task in developing any tool for natural language processing which otherwise hampers the efficiency of the developed systems. The accuracy of the output generated depends on the sense of the context a word is being used in a sentence. Word Sense Disambiguation is used for resolving the ambiguous words. This paper presents a brief introduction to Word Sense Disambiguation, an overview of the approaches that can be used solving ambiguity and work done for various Indian languages.

Dictionary-based Method. And by using the WordNet, we extract concepts of each of the words and Compare them with each other. And by scoring on each of the concepts of the ambiguous word, we chose the correct concept. Keywords: Word Sense Disambiguation, Supervised Method, Unsupervised Method, Semi-Supervised Method, Knowledge base and Dictionary-based Method, Lesk Algorithm.

Proceedings of the 2nd workshop on Learning language in logic and the 4th conference on Computational natural language learning -, 2000

This paper describes a set of comparative experiments, including cross-corpus evaluation, between five alternative algorithms for supervised Word Sense Disambiguation (WSD), namely Naive Bayes, Exemplar-based learning, SNoW, Decision Lists, and Boosting. Two main conclusions can be drawn: 1) The LazyBoosting algorithm outperforms the other four state-of-theart algorithms in terms of accuracy and ability to tune to new domains; 2) The domain dependence of WSD systems seems very strong and suggests that some kind of adaptation or tuning is required for cross-corpus application.