Information Extraction Techniques Research Articles

Assessment of Information Extraction Techniques, Models and Systems

The present article aims to review and evaluate the practiced and classical techniques, tools, models, and systems concerning automatic information extraction (IE) from published scientific documents like research articles, patents, theses, technical reports, and case studies etc. IE is performed for various reasons such as better indexing, archiving, searching, and retrieving. That is mainly used by the search engines and the indexing services as well the digital libraries and semantic web. In this regard, several studies have been conducted targeting various nature of documents. The study pays special consideration to the successful IE models, algorithms and approaches applied to structural IE from published documents. To grasp this, the paper is classified into several segments and each segment covers a significant aspect of IE. Furthermore, to validate their benefits and drawbacks, a comparative study of all the approaches have been conducted in terms of various performance factors like precision, accuracy, recall and F-score. Potential areas of improvement have been emphasized as research gap for the scholars in the closely related areas. Ultimately, a comprehensive summary of the evaluation is presented in tabular form and review is concluded. It was observed that the hybrid methods outperform the other methods due to their versatile nature to address various document formats.

Novel Perspectives for the Management of Multilingual and Multialphabetic Heritages through Automatic Knowledge Extraction: The DigitalMaktaba Approach.

The linguistic and social impact of multiculturalism can no longer be neglected in any sector, creating the urgent need of creating systems and procedures for managing and sharing cultural heritages in both supranational and multi-literate contexts. In order to achieve this goal, text sensing appears to be one of the most crucial research areas. The long-term objective of the DigitalMaktaba project, born from interdisciplinary collaboration between computer scientists, historians, librarians, engineers and linguists, is to establish procedures for the creation, management and cataloguing of archival heritage in non-Latin alphabets. In this paper, we discuss the currently ongoing design of an innovative workflow and tool in the area of text sensing, for the automatic extraction of knowledge and cataloguing of documents written in non-Latin languages (Arabic, Persian and Azerbaijani). The current prototype leverages different OCR, text processing and information extraction techniques in order to provide both a highly accurate extracted text and rich metadata content (including automatically identified cataloguing metadata), overcoming typical limitations of current state of the art approaches. The initial tests provide promising results. The paper includes a discussion of future steps (e.g., AI-based techniques further leveraging the extracted data/metadata and making the system learn from user feedback) and of the many foreseen advantages of this research, both from a technical and a broader cultural-preservation and sharing point of view.

An adaptable, high-performance relation extraction system for complex sentences

The rapid proliferation of text data has lead to an increase in the use of Information Extraction (IE) techniques to automatically extract key information in a fast and effective manner. Relation Extraction (RE), a sub-task of IE focuses on extracting semantic relations from free natural language text and is crucial for further applications including Question Answering, Information Retrieval, Knowledge Base construction, Text Summarization, etc. Literature shows that supervised learning approaches were widely used in RE. However, the performance of supervised methodologies depend on the availability of domain-specific annotated datasets which is not viable for many of the domains including legal, financial, insurance etc. In recent times, Open Information Extraction (OIE) techniques address this issue, by facilitating domain-independent extraction of relations from large text corpora with no demand for domain-specific tagged data and predefined relation classes. Even though OIE systems are fast and simple to implement, they are less effective in handling complex sentences, and often produce redundant extractions.This paper proposes an efficient RE system to extract domain-specific relations from natural language text, consisting of Knowledge-based and Semi-supervised learning systems, integrated with domain ontology. We evaluated the performance of proposed work on ‘judicial domain” as a use case and found that it overcomes the flaws and limitations of existing RE approaches, by achieving better results in terms of precision and recall. On further analysis, we found that the proposed system outperforms existing cutting-edge OIE systems on varying sentence length and complexity.

INFORMATION EXTRACTION USING NATURAL LANGUAGE PROCESSING

The expanding volume of data/Information creates a new challenge for Information Extraction techniques. The amount of unstructured data has increased in recent years. When we use this data in a clean and better way, we can extract wide variety of beneficial outcomes. Reduction of a text to its essential content, is a very complex problem which, despite the progress in the area thus far, poses many challenges to the scientific community. It is also a relevant application in today’s information society given the enormous amount of production and processing of data, leading to exponential growth of textual information online and the need to momentarily assess the contents of text collections. Information Extraction using Natural Language Text is used to process, or extract and encode information from unstructured data and produce desired output data according to the application. This review article is about how Information Extraction is done using NLP (Natural language Processing)

Entropy Estimators in SAR Image Classification.

Remotely sensed data are essential for understanding environmental dynamics, for their forecasting, and for early detection of disasters. Microwave remote sensing sensors complement the information provided by observations in the optical spectrum, with the advantage of being less sensitive to adverse atmospherical conditions and of carrying their own source of illumination. On the one hand, new generations and constellations of Synthetic Aperture Radar (SAR) sensors provide images with high spatial and temporal resolution and excellent coverage. On the other hand, SAR images suffer from speckle noise and need specific models and information extraction techniques. In this sense, the family of distributions is a suitable model for SAR intensity data because it describes well areas with different degrees of texture. Information theory has gained a place in signal and image processing for parameter estimation and feature extraction. Entropy stands out as one of the most expressive features in this realm. We evaluate the performance of several parametric and non-parametric Shannon entropy estimators as input for supervised and unsupervised classification algorithms. We also propose a methodology for fine-tuning non-parametric entropy estimators. Finally, we apply these techniques to actual data.

Frequency information extraction based on time-frequency ridges for characterizing thin sand bodies

Abstract Time-frequency analysis (TFA) is a widely used reservoir prediction technology. During the identification of sand bodies using TFA, abnormally high values in the high-frequency portion of the spectrum usually indicate the presence of thin sand bodies. However, owing to the complexity of the depositional environment, the spatial distribution of thin sand bodies is often variable, and multiple phases of sand bodies—superimposed on each other—can occur with uneven thicknesses. A frequency information extraction technique based on time-frequency ridges (TFRs) is proposed in this study to identify stable TFRs in the time-frequency spectrum of a rearranged smooth pseudo-Wigner–Ville distribution with high energy aggregation, as well as to extract the main- and high-frequency energy contents relative to them. To enhance the ability of these energy components to detect thin sand bodies, a frequency-energy weighting factor is computed. By performing the aforementioned steps, information about each frequency band can be obtained and a comprehensive depiction of sand bodies with different thicknesses achieved. It is concluded that by using the proposed method thin sand bodies such as small riverbeds can be effectively characterized using seismic data.

Object Detection Method for High Resolution Remote Sensing Imagery Based on Convolutional Neural Networks with Optimal Object Anchor Scales

ABSTRACT Object detection is a key technique for automatic information extraction, analysis and understanding of high spatial resolution remote-sensing images (HSRIs). Object anchor scales are a critical factor for object detection of HSRIs based on convolutional neural networks (CNNs). With respect to adaptively obtaining optimal object anchor scales for object detection of HSRIs, this paper proposes a novel object detection method for HSRIs based on CNNs with optimal object anchor scales. First, optimal object anchor scales for object detection of HSRIs are obtained using an adaptive object-scale learning operator. Then, a CNN object detection framework for HSRIs is designed based on optimal object anchor scales. Using multiple-object detection datasets, the proposed method is compared with some state-of-the-art object detection algorithms. Experimental results show that the proposed method can achieve 90.42%, 91.98% and 85.07% mAP on WHU-RSONE, UCAS-AOD and HSRC2016, respectively, and outperforms state-of-the-art object detection algorithms.

A Novel Approach to Automate Complex Software Modularization Using a Fact Extraction System

Complex software systems that support organizations are updated regularly, which can erode system architectures. Moreover, documentation is rarely synchronized with the changes to the software system. This creates a slew of issues for future software maintenance. To this goal, information extraction tools use exact approaches to extract entities and their corresponding relationships from source code. Such exact approaches extract all features, including those that are less prominent and may not be significant for modularization. In order to resolve the issue, this work proposes an enhanced approximate information extraction approach, namely, fact extractor system for Java applications (FESJA) that aims to automate software modularization using a fact extraction system. The proposed FESJA technique extracts all the entities along with their corresponding more dominant formal and informal relationships from a Java source code. Results demonstrate the improved performance of FESJA, by extracting 74 (classes), 43 (interfaces), and 31 (enumeration), in comparison with eminent information extraction techniques.

Big Data Management in Drug–Drug Interaction: A Modern Deep Learning Approach for Smart Healthcare

The detection and classification of drug–drug interactions (DDI) from existing data are of high importance because recent reports show that DDIs are among the major causes of hospital-acquired conditions and readmissions and are also necessary for smart healthcare. Therefore, to avoid adverse drug interactions, it is necessary to have an up-to-date knowledge of DDIs. This knowledge could be extracted by applying text-processing techniques to the medical literature published in the form of ‘Big Data’ because, whenever a drug interaction is investigated, it is typically reported and published in healthcare and clinical pharmacology journals. However, it is crucial to automate the extraction of the interactions taking place between drugs because the medical literature is being published in immense volumes, and it is impossible for healthcare professionals to read and collect all of the investigated DDI reports from these Big Data. To avoid this time-consuming procedure, the Information Extraction (IE) and Relationship Extraction (RE) techniques that have been studied in depth in Natural Language Processing (NLP) could be very promising. Since 2011, a lot of research has been reported in this particular area, and there are many approaches that have been implemented that can also be applied to biomedical texts to extract DDI-related information. A benchmark corpus is also publicly available for the advancement of DDI extraction tasks. The current state-of-the-art implementations for extracting DDIs from biomedical texts has employed Support Vector Machines (SVM) or other machine learning methods that work on manually defined features and that might be the cause of the low precision and recall that have been achieved in this domain so far. Modern deep learning techniques have also been applied for the automatic extraction of DDIs from the scientific literature and have proven to be very promising for the advancement of DDI extraction tasks. As such, it is pertinent to investigate deep learning techniques for the extraction and classification of DDIs in order for them to be used in the smart healthcare domain. We proposed a deep neural network-based method (SEV-DDI: Severity-Drug–Drug Interaction) with some further-integrated units/layers to achieve higher precision and accuracy. After successfully outperforming other methods in the DDI classification task, we moved a step further and utilized the methods in a sentiment analysis task to investigate the severity of an interaction. The ability to determine the severity of a DDI will be very helpful for clinical decision support systems in making more accurate and informed decisions, ensuring the safety of the patients.

Review of Semantic Importance and Role of using Ontologies in Web Information Retrieval Techniques

The Web contains an enormous amount of information, which is managed to accumulate, researched, and regularly used by many users. The nature of the Web is multilingual and growing very fast with its diverse nature of data including unstructured or semi-structured data such as Websites, texts, journals, and files. Obtaining critical relevant data from such vast data with its diverse nature has been a monotonous and challenging task. Simple key phrase data gathering systems rely heavily on statistics, resulting in a word incompatibility problem related to a specific word's inescapable semantic and situation variants. As a result, there is an urgent need to arrange such colossal data systematically to find out the relevant information that can be quickly analyzed and fulfill the users' needs in the relevant context. Over the years ontologies are widely used in the semantic Web to contain unorganized information systematic and structured manner. Still, they have also significantly enhanced the efficiency of various information recovery approaches. Ontological information gathering systems recover files focused on the semantic relation of the search request and the searchable information. This paper examines contemporary ontology-based information extraction techniques for texts, interactive media, and multilingual data types. Moreover, the study tried to compare and classify the most significant developments utilized in the search and retrieval techniques and their major disadvantages and benefits.

Profiled side channel attacks against the RSA cryptosystem using neural networks

Profiled attacks employing machine learning techniques for secret information extraction are currently one of the main interests of the research community working on side channel attacks. In this work, we tackle the use of machine learning methods to perform a power consumption based side channel attack against asymmetric cryptosystems. In particular, we analyze different machine-learning based strategies to retrieve the secret exponent from a square-and-multiply-always modular exponentiation in RSA cryptosystem software implementations. We validate our proposals conducting effective attacks against the modular exponentiation contained in the BearSSL software library, a timing side-channel hardened public library. The experimental validation shows that our attack strategy improves the state-of-art solution and is able to retrieve the correct exponent value with a single trace from the attacked device, and, in the worst case, with an exhaustive search in a set of candidates containing a few hundred possible values for the sought secret.

Strategies for Practical Hybrid Attack Graph Generation and Analysis

As an analytical tool in cyber-security, an attack graph (AG) is capable of discovering multi-stage attack vectors on target computer networks. Cyber-physical systems (CPSs) comprise a special type of network that not only contains computing devices but also integrates components that operate in the continuous domain, such as sensors and actuators. Using AGs on CPSs requires that the system models and exploit patterns capture both token- and real-valued information. In this article, we describe a hybrid AG model for security analysis of CPSs and computer networks. Specifically, we focus on two issues related to applying the model in practice: efficient hybrid AG generation and techniques for information extraction from them. To address the first issue, we present an accelerated hybrid AG generator that employs parallel programming and high performance computing (HPC). We conduct performance tests on CPU and GPU platforms to characterize the efficiency of our parallel algorithms. To address the second issue, we introduce an analytical regimen based on centrality analysis and apply it to a hybrid AG generated for a target CPS system to discover effective vulnerability remediation solutions.

Optimal dynamic treatment regime estimation using information extraction from unstructured clinical text.

The wide-scale adoption of electronic health records (EHRs) provides extensive information to support precision medicine and personalized health care. In addition to structured EHRs, we leverage free-text clinical information extraction (IE) techniques to estimate optimal dynamic treatment regimes (DTRs), a sequence of decision rules that dictate how to individualize treatments to patients based on treatment and covariate history. The proposed IE of patient characteristics closely resembles "The clinical Text Analysis and Knowledge Extraction System" and employs named entity recognition, boundary detection, and negation annotation. It also utilizes regular expressions to extract numerical information. Combining the proposed IE with optimal DTR estimation, we extract derived patient characteristics and use tree-based reinforcement learning (T-RL) to estimate multistage optimal DTRs. IE significantly improved the estimation in counterfactual outcome models compared to using structured EHR data alone, which often include incomplete data, data entry errors, and other potentially unobserved risk factors. Moreover, including IE in optimal DTR estimation provides larger study cohorts and a broader pool of candidate tailoring variables. We demonstrate the performance of our proposed method via simulations and an application using clinical records to guide blood pressure control treatments among critically ill patients with severe acute hypertension. This joint estimation approach improves the accuracy of identifying the optimal treatment sequence by 14-24% compared to traditional inference without using IE, based on our simulations over various scenarios. In the blood pressure control application, we successfully extracted significant blood pressure predictors that are unobserved or partially missing from structuredEHR.

Using the Pauli Scattering Mechanisms for Analysis of Polarimetric SAR Images from RADARSAT Constellation and TanDEM-X Missions

This work describes techniques for extraction of information in polarimetric images from RADARSAT constellation mission (RCM) and TanDEM-X (TDX) and TerraSAR-X (TSX) satellites. Using the boundary conditions of Maxwell's equations it is shown how the single bounce and double bounce scattering can be extracted from quad-pol and compact polarimetric (CP) SAR. We develop new equations to calculate the Pauli decomposition from simulated CP data using RCM quad-pol data. The RCM CP option is circular transmit, linear receive (CTLR). We find that using quad-pol data performs better than using CP data for Pauli decomposition, especially extraction of the HV component. Estimation of single bounce and double bounce scattering from RCM CP data is feasible and yields more information about scattering properties than the CP components alone. We show how more information can be extracted from bistatic TSX/TDX quad-pol data by estimation of the coherence of the different Pauli components and optimization of the interferometric coherence using ships as detection objects. This is a unique opportunity to use optimized coherence on moving ships for characterizing their structure.

Urdu Wikification and Its Application in Urdu News Recommendation System

Wikification is the process of linking the entities found in a sample text to their individual Wikipedia or Wikidata pages. Many natural language processing applications, including question-answering systems, information retrieval, fraud detection, and recommendation systems(RS), can benefit from this information extraction technique. There has been a great deal of effort put towards entity-linking(EL) for both Asian and Western languages, with several datasets and numerous proposed methodologies. Despite millions of Urdu language users globally, the entity-linking for the Urdu language has not been worked on, to the best of our knowledge. This work proposes an Urdu EL pipeline to identify named entities in text and link them to Wikidata. Secondly, a dataset of 550 Urdu news titles relating to their respective Wiki-ids has been prepared for the examination. Third, utilizing the proposed EL pipeline, 16738 news articles from the first-ever Urdu news RS dataset of 100 users are annotated. Fourthly, a sub Knowledge graph (KG) of 8439 entities and 23080 relationship tuples is retrieved from Wikidata. The Trans-E algorithm is then used to create KG embeddings so that the extracted KG may be used in an Urdu news RS. The final accuracy of Urdu news RS is 60.8%.

News Monitor: A Framework for Exploring News in Real-Time

News articles generated by online media are a major source of information. In this work, we present News Monitor, a framework that automatically collects news articles from a wide variety of online news portals and performs various analysis tasks. The framework initially identifies fresh news (first stories) and clusters articles about the same incidents. For every story, at first, it extracts all of the corresponding triples and, then, it creates a knowledge base (KB) using open information extraction techniques. This knowledge base is then used to create a summary for the user. News Monitor allows for the users to use it as a search engine, ask their questions in their natural language and receive answers that have been created by the state-of-the-art framework BERT. In addition, News Monitor crawls the Twitter stream using a dynamic set of “trending” keywords in order to retrieve all messages relevant to the news. The framework is distributed, online and performs analysis in real-time. According to the evaluation results, the fake news detection techniques utilized by News Monitor allow for a F-measure of 82% in the rumor identification task and an accuracy of 92% in the stance detection tasks. The major contribution of this work can be summarized as a novel real-time and scalable architecture that combines various effective techniques under a news analysis framework.

From Tokenization to Self-Supervision: Building a High-Performance Information Extraction System for Chemical Reactions in Patents.

Chemical reactions and experimental conditions are fundamental information for chemical research and pharmaceutical applications. However, the latest information of chemical reactions is usually embedded in the free text of patents. The rapidly accumulating chemical patents urge automatic tools based on natural language processing (NLP) techniques for efficient and accurate information extraction. This work describes the participation of the Melax Tech team in the CLEF 2020—ChEMU Task of Chemical Reaction Extraction from Patent. The task consisted of two subtasks: (1) named entity recognition to identify compounds and different semantic roles in the chemical reaction and (2) event extraction to identify event triggers of chemical reaction and their relations with the semantic roles recognized in subtask 1. To build an end-to-end system with high performance, multiple strategies tailored to chemical patents were applied and evaluated, ranging from optimizing the tokenization, pre-training patent language models based on self-supervision, to domain knowledge-based rules. Our hybrid approaches combining different strategies achieved state-of-the-art results in both subtasks, with the top-ranked F1 of 0.957 for entity recognition and the top-ranked F1 of 0.9536 for event extraction, indicating that the proposed approaches are promising.

An overview of information extraction techniques for legal document analysis and processing

<span>In an Indian law system, different courts publish their legal proceedings every month for future reference of legal experts and common people. Extensive manual labor and time are required to analyze and process the information stored in these lengthy complex legal documents. Automatic legal document processing is the solution to overcome drawbacks of manual processing and will be very helpful to the common man for a better understanding of a legal domain. In this paper, we are exploring the recent advances in the field of legal text processing and provide a comparative analysis of approaches used for it. In this work, we have divided the approaches into three classes NLP based, deep learning-based and, KBP based approaches. We have put special emphasis on the KBP approach as we strongly believe that this approach can handle the complexities of the legal domain well. We finally discuss some of the possible future research directions for legal document analysis and processing.</span>

Extraction of Frequency Information for the Reliable Screening of Rotor Electrical Faults Via Torque Monitoring in Induction Motors

Conventional diagnostic methods applied to industrial induction motors, such as the Motor Current Signature Analysis, may lead to false-negative diagnostic outcomes in several cases. Such a case consists of the non-adjacent rotor breakages occurrence. Various alternatives with advanced digital signal processing algorithms have been proposed that concern the monitoring and analysis of the stator current, or the stray magnetic flux, of the motor during the transient start-up. Those methods work efficiently in most cases; however, the real issue is that most large industrial motors have very few start-ups during their long operating life time. In that sense, it is not feasible to implement the transient analysis-based methods. This paper addresses an alternative methodology that solves this issue for induction motors at steady state. The method relies on a two-stage signal processing technique for frequency information tracking and extraction, where the higher harmonic index of the motors torque around the sixth harmonic is evaluated during each stage. By the results of the method, it is evident that the fault and its severity level can be reliably detected at the steady state. The methods efficacy is proven valid even for challenging cases of large industrial motors, where the likelihood of a false diagnostic decision is increased during the signature screening.

A Natural Language Processing Approach to Understanding Context in the Extraction and GeoCoding of Historical Floods, Storms, and Adaptation Measures

Despite the known financial, economical, and humanitarian impacts of hurricanes and the floods that follow, datasets consisting of flood and flood risk reduction projects are either small in scope, lack in details, or held privately by commercial holders. However, with the amount of online data growing exponentially, we have seen a rise of information extraction techniques on unstructured text to drive insights. On one hand, social media in particular has seen a tremendous increase in popularity. On the other hand, despite this popularity, social media has proven to be unreliable and difficult to extract full information from. In contrast, online newspapers are often vetted by a journalist, and consist of more fine details. As a result, in this paper we leverage Natural Language Processing (NLP) to create a hybrid Named-Entity Recognition (NER) model that employs a domain-specific machine learning model, linguistic features, and rule-based matching to extract information from newspapers. To the knowledge of the authors, this model is the first of its kind to extract detailed flooding information and risk reduction projects over the entire contiguous United States. The approach used in this paper expands upon previous similar works by widening the geographical location and applying techniques to extract information over large documents, with minimal accuracy loss from the previous methods. Specifically, our model is able to extract information such as street closures, project costs, and metrics. Our validation indicates an F1 score of 72.13% for the NER model entity extraction, a binary classification location filter with a score of 73%, and an overall performance only 8.4% lower than a human validator against a gold-standard. Through this process, we find the location of 27,444 streets, 181,076 flood risk reduction projects, and 435,353 storm locations throughout the United States in the past two decades.