Vocabulary Mismatch Research Articles

SKYPER: Legal case retrieval via skeleton-aware hypergraph embedding in the hyperbolic space

Legal Case Retrieval (LCR) plays a significant role in ensuring judicial justice in various legal systems and has recently received increasing attention in Information Retrieval (IR) research. Existing LCR methods typically encode documents into low-dimensional vectors by using either bag of keywords or pretrained language models. However, such approaches generally overlook the interactions among numerous cases and the hierarchy skeleton between event and event type. In this paper, we propose to construct an event-aware knowledge hypergraph for representing legal cases. And we put forward a SKeleton-aware hYPERgraph representation framework named SKYPER, to learn the case embeddings. Concretely, a global skeleton aggregates coarse-grained type information and roughly locates a case. A local skeleton learns the fine-grained information and precisely locates a case. And SKYPER also uses hyperbolic space strengthens the hierarchy between event and event type and avoids vocabulary mismatch. At the retrieval stage, the SKYPER learned case embedding is leveraged for further case similarity matching with a few annotated similarity ranking data. Extensive experiments demonstrate that SKYPER outperforms existing models on the public LCR datasets LeCard and COLIEE20-T1 in both supervised and unsupervised settings. Especially, the retrieval efficiency is improved by a large margin.

Document retrieval using clustering-based Aquila hash-Q optimization with query expansion based on pseudo relevance feedback

A document retrieval system helps users to retrieve the relevant documents corresponding to their query quickly and easily. In the real world, document retrieval is a difficult task due to high volumes of data, unstructured data, and different formats of data. Even though many research techniques are introduced, major problems like vocabulary mismatch and non-linear matching still need to be solved. In this work, the Aquila hash-q optimizer is the proposed matching technique with the clustering technique to retrieve the document in a time-efficient manner for the user query without collision. First, preprocessing is done by eliminating the stop words from the document, stemming, and grouping documents in a cluster into a single document using Hierarchical Density-based Sampling Spatial Cluster of Applications with Noise (HDBSSCAN) clustering. This clustering algorithm is powerful, robust to noise, and scalable and identifies clusters of documents that are related to each other. Additionally, the sampling technique used in this clustering algorithm increases the clustering speed by reducing the size of the document which improves the performance of document retrieval systems. Secondly, the queries are searched using the Aquila hash-q optimizer matching technique by which the relevant documents are retrieved. The Aquila hash-q optimization works by pre-computing a hash table of the terms in a document collection and then using this hash table to quickly identify the relevant documents from the given query. This can significantly improve the speed of document retrieval, especially for large document collections. Aquila hash-q optimization can improve the accuracy, efficiency, and scalability of document retrieval systems. The effectiveness of the Hierarchical Density-Based Clustering Aquila Optimization approach is determined by various analyses through NPL, LISA, and CACM data in terms of precision @ 5 (0.497), precision @ 10 (0.425), Mean Average Precision (MAP) (0.462) by comparing our approach with various methods. As a result, the Aquila hash-q optimizer is the proposed matching technique to retrieve the document in a time-efficient manner for the user query without collision.

Systematic Literature Review on Ontology-based Indonesian Question Answering System

Question-Answering (QA) systems at the intersection of natural language processing, information retrieval, and knowledge representation aim to provide efficient responses to natural language queries. These systems have seen extensive development in English and languages like Indonesian present unique challenges and opportunities. This literature review paper delves into the state of ontology-based Indonesian QA systems, highlighting critical challenges. The first challenge lies in sentence understanding, variations, and complexity. Most systems rely on syntactic analysis and struggle to grasp sentence semantics. Complex sentences, especially in Indonesian, pose difficulties in parsing, semantic interpretation, and knowledge extraction. Addressing these linguistic intricacies is pivotal for accurate responses. Secondly, template-based SPARQL query construction, commonly used in Indonesian QA systems, suffers from semantic gaps and inflexibility. Advanced techniques like semantic matching algorithms and dynamic template generation can bridge these gaps and adapt to evolving ontologies. Thirdly, lexical gaps and ambiguity hinder QA systems. Bridging vocabulary mismatches between user queries and ontology labels remains a challenge. Strategies like synonym expansion, word embedding, and ontology enrichment must be explored further to overcome these challenges. Lastly, the review discusses the potential of developing multi-domain ontologies to broaden the knowledge coverage of QA systems. While this presents complex linguistic and ontological challenges, it offers the advantage of responding to various user queries across various domains. This literature review identifies crucial challenges in developing ontology-based Indonesian QA systems and suggests innovative approaches to address these challenges.

A Systematic Review of Automated Query Reformulations in Source Code Search

Fixing software bugs and adding new features are two of the major maintenance tasks. Software bugs and features are reported as change requests. Developers consult these requests and often choose a few keywords from them as an ad hoc query. Then they execute the query with a search engine to find the exact locations within software code that need to be changed. Unfortunately, even experienced developers often fail to choose appropriate queries, which leads to costly trials and errors during a code search. Over the years, many studies have attempted to reformulate the ad hoc queries from developers to support them. In this systematic literature review, we carefully select 70 primary studies on query reformulations from 2,970 candidate studies, perform an in-depth qualitative analysis (e.g., Grounded Theory), and then answer seven research questions with major findings. First, to date, eight major methodologies (e.g., term weighting, term co-occurrence analysis, thesaurus lookup) have been adopted to reformulate queries. Second, the existing studies suffer from several major limitations (e.g., lack of generalizability, the vocabulary mismatch problem, subjective bias) that might prevent their wide adoption. Finally, we discuss the best practices and future opportunities to advance the state of research in search query reformulations.

Bug Localization with Features Crossing and Structured Semantic Information Matching

Bug localization techniques aim to locate the relevant buggy source files according to the bug described by the given bug report, so as to improve the localization efficiency of developers and reduce the cost of software maintenance. The traditional bug localization techniques based on Information Retrieval (IR) usually use the classical text retrieval model and combines the specific domain knowledge features in software engineering to locate the bugs. However, there exists the vocabulary mismatch problem between the bug report and the source file, which may affect the performance of bug localization. Therefore, the relevant deep learning model was introduced later to compute the similarity between the bug report and the source file from the perspective of semantic features. Bug localization approaches based on IR and deep learning have their own advantages and disadvantages, so this paper proposes a model named LocFront which combines IR and deep learning. On the one hand, the Features Crossing module in LocFront carries out the deep crossing operation on the extracted software-specific features to fully acquire the linear and nonlinear relationships. On the other hand, the Structured Semantic Information Matching module in LocFront performs semantic matching on the structured information between the bug report and the source file. Then the Fusion module in LocFront fuses the results of the two above modules to obtain the final localization score. The experimental results on five benchmark datasets show that LocFront outperforms the state-of-the-art bug localization approaches.

Learning to Ask: Conversational Product Search via Representation Learning

Online shopping platforms, such as Amazon and AliExpress, are increasingly prevalent in society, helping customers purchase products conveniently. With recent progress in natural language processing, researchers and practitioners shift their focus from traditional product search to conversational product search. Conversational product search enables user-machine conversations and through them collects explicit user feedback that allows to actively clarify the users’ product preferences. Therefore, prospective research on an intelligent shopping assistant via conversations is indispensable. Existing publications on conversational product search either model conversations independently from users, queries, and products or lead to a vocabulary mismatch. In this work, we propose a new conversational product search model, ConvPS, to assist users in locating desirable items. The model is first trained to jointly learn the semantic representations of user, query, item, and conversation via a unified generative framework. After learning these representations, they are integrated to retrieve the target items in the latent semantic space. Meanwhile, we propose a set of greedy and explore-exploit strategies to learn to ask the user a sequence of high-performance questions for conversations. Our proposed ConvPS model can naturally integrate the representation learning of the user, query, item, and conversation into a unified generative framework, which provides a promising avenue for constructing accurate and robust conversational product search systems that are flexible and adaptive. Experimental results demonstrate that our ConvPS model significantly outperforms state-of-the-art baselines.

Aspect-Bas1ed Sentiment Analysis on Twitter Using Logistic Regression with FastText Feature Expansion

Social media has recently been widely used by users, especially Indonesians, as a place to express themselves in sentences, pictures, sounds, or videos. Twitter is one of the social media favored by people of diverse ages. Twitter is a social media that provides features like social media in general. However, Twitter has a unique feature where users can send or read text messages limited to only a few characters. Therefore, user tweets with topics related to a particular product can be utilized by companies to become input in the development of these products. This research was conducted using tweet data on the topic of Telkomsel, which is divided into two aspects, namely signal and service. Aspect-based sentiment analysis of Telkomsel was carried out using Logistic Regression with FastText feature expansion to reduce vocabulary mismatch in tweets so that the classification stage can be performed optimally. In addition, the Synthetic Minority Oversampling Technique (SMOTE) sampling method was applied to overcome data imbalance. The test results prove that feature expansion can improve F1-Score values for signal and service aspects. For the signal aspect, F1-Score increased by 3.33% from the baseline with a value of 96.48%. While for the service aspect, F1-Score increased by 12.91% from the baseline with a value of 95.57%.

Empowering the Human as the Fitness Function in Search-Based Model-Driven Engineering

In Search-Based Software Engineering, more than 100 works have involved the human in the search process to obtain better results. However, the case where the human completely replaces the fitness function remains neglected. There is a good reason for that; no matter how intelligent the human is, humans cannot assess millions of candidate solutions as heuristics do. In this work, we study the influence of using the Human as the Fitness Function (HaFF) on the quality of the results. To do that, we focus on Search-Based Model-Driven Engineering (SBMDE) because inspecting models should require less human effort than inspecting code thanks to the abstraction of models. Therefore, we analyze the impact of HaFF in a real-world industrial case study of feature location in models. Furthermore, we also consider a reformulation operation (replacement) in the evaluation because a recent work reported that this operation significantly reduces the number of iterations required in comparison to the widespread crossover and mutation operations. The combination of HaFF and the reformulation operation (HaFF_R) improves the results of the best baseline by 0.15% in recall and 14.26% in precision. Analyzing the results, we learned how to better leverage HaFF_R, which increased the improvement with regard to the best baseline to 1.15% in recall and 20.05% in precision. HaFF_R significantly improves precision because humans are immune to the main limitations of the baselines: vocabulary mismatch and tacit knowledge. A focus group confirmed the acceptance of HaFF. These results are relevant for SBMDE because feature location is one of the main activities performed during maintenance and evolution. Our results, and what we learned from them, can also motivate and help other researchers to explore the benefits of HaFF. In fact, we provide a guideline that further discusses how to apply HaFF to other software engineering problems.

Word2Vec on Sentiment Analysis with Synthetic Minority Oversampling Technique and Boosting Algorithm

Customer opinion is an important aspect in determining the success of a company or service provider. By determining the sentiment of the existing opinion, the company can use it as an evaluation material to improve the quality of the service or product provided. Sentiment analysis can be used as a measure of opinion sentiment with input data in the form of a corpus which will be classified into positive or negative classes to obtain the level of customer satisfaction with a product or service. Aspect-based sentiment analysis can be used by companies to analyze more specifically and find out what aspects need to be improved. In this research, an aspect-based sentiment analysis was conducted on Telkomsel users on Twitter. The data used is 16,992 tweets from users who discuss several aspects such as Telkomsel's services and signals in Twitter. In this research Word2Vec was used for feature expansion to minimize vocabulary mismatch caused by limited words in tweets. The results showed that Word2Vec, Synthetic Minority Oversampling Technique (SMOTE), and Boosting algorithm combination with Logistic Regression classifier achieve highest accuracy of 95.10% for signal aspect and using hyperparameters makes the service aspect get the highest accuracy of 93.34%.

Aspect Based Sentiment Analysis with FastText Feature Expansion and Support Vector Machine Method on Twitter

Social media such as Twitter has now become very close to society. Twitter users can express current issues, their opinions, product reviews, and many other things both positive and negative. Twitter is also used by companies to monitor the assessment of their products among the public as insight that will be used to evaluate what aspects of their products need to be further developed. Twitter with its limitation of only allowing users to post a maximum tweet of 280 characters will make a lot of abbreviated and difficult to understand words used, so it will allow vocabulary mismatch problems to occur. Therefore, in this paper, research conducted on aspect-based sentiment analysis of Telkomsel’s products from the aspects of signal and service by applying feature expansion using Fasttext word embedding to overcome vocabulary mismatch problem and classification with the Support Vector Machine (SVM) method. Sampling technique with Synthetic Minority Oversampling Technique (SMOTE) used to overcome data imbalance. The experimental results show that feature expansion can increase the performance of model. The final results obtained F1-Score value of the model for the signal aspect increased by 27.91% with F1-Score 95.93%, and for the service aspect increased by 42.36% with F1-Score 94.53%.

Feature Expansion Using Word2vec for Hate Speech Detection on Indonesian Twitter with Classification Using SVM and Random Forest

Hate speech is one of the most common cases on Twitter. It is limited to 280 characters in uploading tweets, resulting in many word variations and possible vocabulary mismatches. Therefore, this study aims to overcome these problems and build a hate speech detection system on Indonesian Twitter. This study uses 20,571 tweet data and implements the Feature Expansion method using Word2vec to overcome vocabulary mismatches. Other methods applied are Bag of Word (BOW) and Term Frequency-Inverse Document Frequency (TF-IDF) to represent feature values in tweets. This study examines two methods in the classification process, namely Support Vector Machine (SVM) and Random Forest (RF). The final result shows that the Feature Expansion method with TF-IDF weighting in the Random Forest classification gives the best accuracy result, which is 88,37%. The Feature Expansion method with TF-IDF weighting can increase the accuracy value from several tests in detecting hate speech and overcoming vocabulary mismatches.

Semantic Models for the First-Stage Retrieval: A Comprehensive Review

Multi-stage ranking pipelines have been a practical solution in modern search systems, where the first-stage retrieval is to return a subset of candidate documents and latter stages attempt to re-rank those candidates. Unlike re-ranking stages going through quick technique shifts over the past decades, the first-stage retrieval has long been dominated by classical term-based models. Unfortunately, these models suffer from the vocabulary mismatch problem, which may block re-ranking stages from relevant documents at the very beginning. Therefore, it has been a long-term desire to build semantic models for the first-stage retrieval that can achieve high recall efficiently. Recently, we have witnessed an explosive growth of research interests on the first-stage semantic retrieval models. We believe it is the right time to survey current status, learn from existing methods, and gain some insights for future development. In this article, we describe the current landscape of the first-stage retrieval models under a unified framework to clarify the connection between classical term-based retrieval methods, early semantic retrieval methods, and neural semantic retrieval methods. Moreover, we identify some open challenges and envision some future directions, with the hope of inspiring more research on these important yet less investigated topics.

A hybrid query expansion framework for the optimal retrieval of the biomedical literature

With the proliferation of biomedical literature, it is quite challenging for biomedical scientists to keep them updated with the new advancements. In biomedical literature retrieval systems, the keywords in the user-defined queries are often defined with various lexical variants consequently leading to the vocabulary mismatch (VM). One possible way to cope with these issues is to introduce a query expansion (QE) framework to enrich the original queries with the auxiliary semantically similar terms for each keyword mentioned in a query. In this research, we propose a biomedical QE framework to alleviate the VM. The proposed approach combines the clinical diagnosis information (CDI) and word embeddings (WEs) simultaneously to retrieve the relevant biomedical literature. The process of embeddings vocabulary terms as real-valued and low dimensional vectors referred to as word embedding has garnered significant attention by potentially capturing the implicit semantics. We have exploited threefold word embeddings (Domain-Specific, Domain-Agnostic, and Hybrid) and integrated the embeddings outcomes with the CDI to get the best query combination for the efficient retrieval of biomedical literature. Experimental results procured for the Text REtrieval Conference dataset showed that CDI, when used with the hybrid word embeddings surpassed the WEs trained for the domain-specific and domain-agnostic data. The results demonstrate that the utilization of this unique setup of merging two techniques is a valuable addition to the QE process leading to significantly improved precision rate and VM in biomedical literature retrieval. We hope that our approach would assist investigators to use this query combination to retrieve relevant articles.

Feature Expansion Word2Vec for Sentiment Analysis of Public Policy in Twitter

Social media users, especially on Twitter, can freely express opinions or other information in the form of tweets about anything, including responding to a public policy. In a written tweet, there is a limit of 280 characters per tweet and this allows for problems such as vocabulary mismatches. Therefore, in this study, the feature expansion Word2vec method was applied to overcome when the vocabulary mismatches occur. This study develops and compares the Twitter sentiment analysis system using the feature expansion Word2vec method with the Logistic Regression (LR) and Support Vector Machine (SVM) classification algorithms and the system without the feature expansion Word2Vec method. The results of this study, the feature expansion Word2Vec method on the SVM classification algorithm succeeded in increasing the system accuracy up to 0,99% with an accuracy value of 78,99%.

Analisis Trending Topik Twitter dengan Fitur Ekspansi FastText Menggunakan Metode Logistic Regression

Twitter is a social media that contains information such as the latest news, a person's biography, and tweets from users. Twitter has a feature called trending topics that serves to find out information on certain topics that are currently popular. In fact, it is often difficult to understand what trending topics are happening. Therefore, it is necessary to classify trending topics into a general category. This study aims to analyze and classify Twitter topic trending information by dividing several topic trend labels using the FastText expansion feature method. The FastText expansion feature is used to reduce vocabulary mismatches in a tweet. The classification process of this system will use the Logistic Regression method. The best results were obtained in this study using test data scenarios, 90:10 training data with 76.39% accuracy. The most discussed trending topic from September 2021 to October 2021 was politics with a percentage of 15.83%, followed by religion 12.64% and technology 10.42%

Attention-based skill translation models for expert finding

The growing popularity of community question answering websites can be seen by the growing number of users. Many methods are proposed to identify talented users in these communities, but many of them suffer from vocabulary mismatches. The solution to this problem can be found in translation approaches. The present paper proposes two translation methods for extracting more relevant translations. The proposed methods rely on the attention mechanism. The methods use multi-label classifiers that take each question as input and predict the skills related to the question. Using the attention mechanism, the model is able to focus on specific parts of the given input and predict the correct labels. The ultimate goal of these networks is to predict skills related to questions. Using word attention scores, we can find out how relevant a single word is to a particular skill. As a result of these attention scores, we obtain more relevant translations for each skill. We then use these translations to bridge the lexical gap and improve expert retrieval results. Extensive experiments on two large sub-collections of the StackOverflow dataset demonstrate that the proposed methods outperform the best baseline method by up to 14.11/% MAP improvement.

Hate Speech Detection on Twitter in Indonesia with Feature Expansion Using GloVe

Twitter is one of the popular social media to channel opinions in the form of criticism and suggestions. Criticism could be a form of hate speech if the criticism implies attacking something (an individual, race, or group). With the limit of 280 characters in a tweet, there is often a vocabulary mismatch due to abbreviations which can be solved with word embedding. This study utilizes feature expansion to reduce vocabulary mismatches in hate speech on Twitter containing Indonesian language by using Global Vectors (GloVe). Feature selection related to the best model is carried out using the Logistic Regression (LR), Random Forest (RF), and Artificial Neural Network (ANN) algorithms. The results show that the Random Forest model with 5.000 features and a combination of TF-IDF and Tweet corpus built with GloVe produce the best accuracy rate between the other models with an average of 88,59% accuracy score, which is 1,25% higher than the predetermined Baseline. The number of features used is proven to improve the performance of the system.

PQPS: Prior-Art Query-Based Patent Summarizer Using RBM and Bi-LSTM

A prior-art search on patents ascertains the patentability constraints of the invention through an organized review of prior-art document sources. This search technique poses challenges because of the inherent vocabulary mismatch problem. Manual processing of every retrieved relevant patent in its entirety is a tedious and time-consuming job that demands automated patent summarization for ease of access. This paper employs deep learning models for summarization as they take advantage of the massive dataset present in the patents to improve the summary coherence. This work presents a novel approach of patent summarization named PQPS: prior-art query-based patent summarizer using restricted Boltzmann machine (RBM) and bidirectional long short-term memory (Bi-LSTM) models. The PQPS also addresses the vocabulary mismatch problem through query expansion with knowledge bases such as domain ontology and WordNet. It further enhances the retrieval rate through topic modeling and bibliographic coupling of citations. The experiments analyze various interlinked smart device patent sample sets. The proposed PQPS demonstrates that retrievability increases both in extractive and abstractive summaries.

Implementation Word2Vec for Feature Expansion in Twitter Sentiment Analysis

Abstract 
 Twitter is a microblog-based social media site launched on July 13, 2006. In March 2020, 476.696 tweets about the government policy in COVID-19 spread on Twitter were captured by the Institute for Development of Economics and Finance (Indef). Government policy has a standard meaning, namely a decision systematically made by the government with specific goals and objectives relating to the public interest, whether carried out directly or indirectly. Sentiment analysis analyzes people’s opinions, sentiments, evaluations, attitudes, and emotions from written language. In this decade, Sentiment Analysis is has become a trendy research area. The purpose of this paper is to focus how to implement word2vec using similarity word as a feature expansion for minimize the vocabulary mismatch in Twitter Sentiment Analysis using “word embeddings”. This research contains 11.395 tweets for a dataset, where the dataset will be used in two classifications: Support Vector Machine Algorithm and Artificial Neural Network Algorithm. The output of Word2Vec will be used for feature expansion in this research, where the algorithm of expansion will check in each row in the corpus where has a similarity vector with that word and will replace the word with the similarity of this words if the value is 0. The dataset in Feature Expansion is using 142.545 articles from Indonesian media. The result of this research is ANN is better than SVM, where the ANN without feature expansion gets 68.89 % and using feature expansion gets 72.58 %. For SVM, the final accuracy without feature expansion is 63.95 %, and using feature expansion gets 68.56 %. This research proves that feature expansion can improve the final accuracy.

Word-embedding-based query expansion: Incorporating Deep Averaging Networks in Arabic document retrieval

One of the main issues associated with search engines is the query–document vocabulary mismatch problem, a long-standing problem in Information Retrieval (IR). This problem occurs when a user query does not match the content of stored documents, and it affects most search tasks. Automatic query expansion (AQE) is one of the most common approaches used to address this problem. Various AQE techniques have been proposed; these mainly involve finding synonyms or related words for the query terms. Word embedding (WE) is one of the methods that are currently receiving significant attention. Most of the existing AQE techniques focus on expanding the individual query terms rather the entire query during the expansion process, and this can lead to query drift if poor expansion terms are selected. In this article, we introduce Deep Averaging Networks (DANs), an architecture that feeds the average of the WE vectors produced by the Word2Vec toolkit for the terms in a query through several linear neural network layers. This average vector is assumed to represent the meaning of the query as a whole and can be used to find expansion terms that are relevant to the complete query. We explore the potential of DANs for AQE in Arabic document retrieval. We experiment with using DANs for AQE in the classic probabilistic BM25 model as well as for two recent expansion strategies: Embedding-Based Query Expansion approach (EQE1) and Prospect-Guided Query Expansion Strategy (V2Q). Although DANs did not improve all outcomes when used in the BM25 model, it outperformed all baselines when incorporated into the EQE1 and V2Q expansion strategies.