Text Information Extraction Research Articles

Intelligent Agent for Hurricane Emergency Identification and Text Information Extraction from Streaming Social Media Big Data

This paper presents our research on leveraging social media Big Data and AI to support hurricane disaster emergency response. The current practice of hurricane emergency response for rescue highly relies on emergency call centres. The more recent Hurricane Harvey event reveals the limitations of the current systems. We use Hurricane Harvey and the associated Houston flooding as the motivating scenario to conduct research and develop a prototype as a proof-of-concept of using an intelligent agent as a complementary role to support emergency centres in hurricane emergency response. This intelligent agent is used to collect real-time streaming tweets during a natural disaster event, to identify tweets requesting rescue, to extract key information such as address and associated geocode, and to visualize the extracted information in an interactive map in decision supports. Our experiment shows promising outcomes and the potential application of the research in support of hurricane emergency response.

An Image–Text Dual-Channel Union Network for Person Re-Identification

Multiple people can have similar appearances and portions of images can be occluded or have viewpoint changes in real scenarios, causing the increased difficulty of person re-identification (Re-ID). To address these problems, we propose a dual-channel person Re-ID algorithm that integrates person Re-ID image-text pairs into the classification network for end-to-end learning. We construct an image channel and a text channel, and subsequently extract visual information and text information using a convolutional neural network and a simple recurrent units network, respectively. The text information is used to assist in the learning of visual information, consequently improving the robustness of the visual information. In addition, the visual features are divided into two branches to calculate the global and local features. Global features focus on the overall appearance of a person, whereas local features provide more fine-grained detail. Text information is more accurate and reliable, and it is thus more robust to occlusion and viewpoint changes. Visual information complemented by text information can describe a person more accurately and reliably. Extensive experiments demonstrate our method achieves state-of-the-art performance.

Contrast U-Net driven by sufficient texture extraction for carotid plaque detection.

Ischemic heart disease or stroke caused by the rupture or dislodgement of a carotid plaque poses a huge risk to human health. To obtain accurate information on the carotid plaque characteristics of patients and to assist clinicians in the determination and identification of atherosclerotic areas, which is one significant foundation work. Existing work in this field has not deliberately extracted texture information of carotid from the ultrasound images. However, texture information is a very important part of carotid ultrasound images. To make full use of the texture information in carotid ultrasound images, a novel network based on U-Net called Contrast U-Net is designed in this paper. First, the proposed network mainly relies on a contrast block to extract accurate texture information. Moreover, to make the network better learn the texture information of each channel, the squeeze-and-excitation block is introduced to assist in the jump connection from encoding to decoding. Experimental results from intravascular ultrasound image datasets show that the proposed network can achieve superior performance compared with other popular models in carotid plaque detection.

IMPROVING THE IMAGE-TO-SPEECH SYSTEM ACCURACY THROUGH INTEGRATION OF OPTICAL CHARACTER RECOGNITION AND LANGUAGE PROCESSING TECHNIQUES

Image-to-speech systems are a type of technology allowing for the conversion of visual information, such as images or videos, into auditory output. These systems use complex algorithms and machine learning techniques to recognize and describe visual content, allowing individuals who are visually impaired or blind to access in-formation that would otherwise be inaccessible to them. Image-to-speech systems are becoming increasingly sophisticated and can be integrated into a variety of devices, from smartphones to smart glasses. This article presents an approach to improving the accuracy of the image-to-speech system by incorporating multiple techniques. The proposed system begins by using Tesseract, an optical character recognition (OCR) engine, to extract text infor-mation from images. However, OCR is often imperfect and produces errors, which can impact the accuracy of image-to-speech models. To address this issue, the Text-Davinci-002 engine was applied for post-processing OCR output, which can help to correct errors and improve the accuracy of the extracted text. Finally, the Microsoft Speech API was employed in order to generate speech from the extracted text. By integrating these three techniques, image-to-speech system accuracy was significantly improved. An example of the generated synthetic dataset showed that the proposed techniques improve image-to-speech system accuracy both on word and character levels, and also perform punctuation error correction. This approach can be useful in various applications, including reading text from images, translating written text to speech, and assisting people with visual im-pairments.

Деструктивная коммуникация молодёжи в сети Интернет: социально-политический контекст

The paper presents the general results of the study of destructive behaviour of young people in web communities. At the first stages of the study the aim was to identify destructive communication practices among young people on the Internet, including in the context of political discourse. At the final stage the aim was to identify strategies and techniques in the news content that can form and provoke destructive patterns of users’ responses to various socio-political phenomena (events, decisions and persons associated with these events).
 The study included a total of 530 respondents — representatives of student youth from va­rious universities in Tomsk. The empirical material of the study also included regional news reports on socio-political topics, posted in free access by various communities on VK social network (in Tomsk and Novosibirsk). We analyzed 3,786 posts from Novosibirsk and 887 from Tomsk. For data analysis the authors used qualitative method (content analysis, topic analysis, psycholinguistic analysis, psychological analysis, machine learning methods) and quantitative method (API-based data mining algorithm, digital footprint data collection using data platform (https://lk.data.tsu.ru/), natural language processing methods for text analysis and information extraction, machine learning methods, PolyAnalyst text analysis platform). Within the scope of the project the authors studied the specifics of perception and consumption of political content on social media and identified destructive behaviour among young people in web communities.

Supply chain disruptions and labor shortages: COVID in perspective

We use a neural network language model and dependency parsing to extract textual information about shortages from US corporate filings. This allows us to study the COVID pandemic in the broader context of other macroeconomic environments over the past 25 years. Shortages are much more prevalent during the pandemic than they are during other times, especially in the goods sector and for intermediate inputs. Firms also report significantly more increases in costs. The largest number of shortages occurs at the very end of our sample, in the first half of 2022. Our data are consistent with several known economic facts that are unrelated to the pandemic. • We use a neural network language model and dependency parsing to extract textual information about shortages from US corporate filings. • Shortages are much more prevalent during the pandemic than they are during other times. • This is most pronounced in the goods sector and for intermediate inputs. • Shortages are most prevalent at the very end of our sample, in the first half of 2022. • Our data are consistent with several known economic facts that are unrelated to the pandemic.

Neural entity alignment with cross-modal supervision

The majority of currently available entity alignment (EA) solutions primarily rely on structural information to align entities, which is biased and disregards additional multi-source information. To compensate for inadequate structural details, this article suggests the SKEA framework, which is a simple but flexible framework for Entity Alignment with cross-modal supervision of Supporting Knowledge. We employ a relational aggregate network to specifically utilize the details about the entity and its neighbors. To overcome the limitations of relational features, two multi-modal encode modules are being used to extract visual and textural information. A new set of potential aligned entity pairs are generated by SKEA in each iteration using the knowledge of two reference modalities, which can enhance the model’s supervision. It is important to note that the supporting information used in our framework does not participate in the network’s backpropagation, which considerably improves efficiency and differs dramatically from earlier work. In comparison to existing baselines, experiments demonstrate that our proposed framework can incorporate multi-aspect information efficiently and enable supervisory signals from other modalities to transmit to entities. The maximum performance improvement of 5.24% indicates our suggested framework’s superiority, especially for sparse KGs.

A Novel Named Entity Recognition Algorithm for Hot Strip Rolling Based on BERT-Imseq2seq-CRF Model

Named entity recognition is not only the first step of text information extraction, but also the key process of constructing domain knowledge graphs. In view of the large amount of text data, complex process flow and urgent application needs in the hot strip rolling process, a novel named entity recognition algorithm based on BERT-Imseq2seq-CRF model is proposed in this paper. Firstly, the algorithm uses the BERT preprocessing language model to mine the dependencies in the domain text and obtain the corresponding representation vector. Then, the representation vector is sent to the encoder layer, and the output vector is input to the decoder at the same time, on the premise that the original model only considers the semantic vector. The Teacher-Forcing mechanism is integrated into the decoder layer to randomly modify the labeling results, and error accumulation is avoided to guarantee the sequence recognition effect. Finally, the validity of the labeling results is checked according to the conditional random field constraints, and the overall labeling quality of the algorithm is improved. The experimental results show that this model can efficiently and accurately predict the physical label of hot strip rolling, and the model performance index is better than other models, with the F1-Score reaching 91.47%. This model further provides technical support for information extraction and domain knowledge graph construction of hot strip rolling.

Detection of small yellow croaker freshness by hyperspectral imaging

Rapid discrimination of fresh and repeatedly thawed small yellow croaker is of great significance for monitoring quality and ensuring consumer safety. In this study, we selected 160 fish samples of similar size and divided them equally into four groups (fresh, freeze-thaw once, freeze-thaw twice and freeze-thaw three times). Hyperspectral images were acquired from fresh and repeatedly frozen-thawed fish samples. The grey scale co-occurrence matrix (GLCM) was then applied to extract texture information from the first three principal component (PC) images, and a library for support vector machines (LIBSVM) were employed to discriminate fresh and repeatedly frozen-thawed fish samples using spectral characteristics, texture features and their fusion, respectively. The results indicated that LIBSVM model using the fused data showed the highest classification accuracy of 96.88%, and freshness degradation of fish samples after three freeze-thaw cycles was observed in the second PC image. In the freshness level for validation model, PLSR model achieved good performance withRV2= 0.90 and RPD= 3.17. The current findings demonstrate that hyperspectral imaging(HSI) is feasible for non-destructive determination of small yellow croaker freshness, providing technical guidance for the storage and marketing of aquatic products.

CAW: A Remote-Sensing Scene Classification Network Aided by Local Window Attention.

Remote-sensing image scene data contain a large number of scene images with different scales. Traditional scene classification algorithms based on convolutional neural networks are difficult to extract complex spatial distribution and texture information in images, resulting in poor classification results. In response to the above problems, we introduce the vision transformer network structure with strong global modeling ability into the remote-sensing image scene classification task. In this paper, the parallel network structure of the local-window self-attention mechanism and the equivalent large convolution kernel is used to realize the spatial-channel modeling of the network so that the network has better local and global feature extraction performance. Experiments on the RSSCN7 dataset and the WHU-RS19 dataset show that the proposed network can improve the accuracy of scene classification. At the same time, the effectiveness of the network structure in remote-sensing image classification tasks is verified through ablation experiments, confusion matrix, and heat map results comparison.

Factors affecting text mining based stock prediction: Text feature representations, machine learning models, and news platforms

Text mining techniques have demonstrated their effectiveness for stock market prediction and different text feature representation approaches, (e.g., TF–IDF and word embedding), have been adapted to extract textual information from financial news sources. In addition, different machine learning techniques including deep learning have been employed to construct the prediction models. Various combinations of text feature representations and learning models have been applied for stock prediction, but it is unknown which performs the best or which ones can be regarded as the representative baselines for future research. Moreover, since the textual contents in the financial news articles published on different news platforms are somewhat different, the effect of using different news platforms may have an impact on prediction performance so this is also examined in the experiments comparing eight different combinations comprised of two context-free and two contextualized text feature representations, i.e. TF–IDF, Word2vec, ELMo, and BERT, and three learning techniques, i.e. SVM, CNN, and LSTM. The experimental results show that CNN+Word2vec and CNN+BERT perform the best. The textual material is taken from three public news platforms including Reuters, CNBC, and The Motley Fool. We found that the learning models constructed and the news platforms used can certainly affect the prediction of stock prices between different companies.

Detection of Depression Severity Using Bengali Social Media Posts on Mental Health: Study Using Natural Language Processing Techniques.

BackgroundThere are a myriad of language cues that indicate depression in written texts, and natural language processing (NLP) researchers have proven the ability of machine learning and deep learning approaches to detect these cues. However, to date, these approaches bridging NLP and the domain of mental health for Bengali literature are not comprehensive. The Bengali-speaking population can express emotions in their native language in greater detail.ObjectiveOur goal is to detect the severity of depression using Bengali texts by generating a novel Bengali corpus of depressive posts. We collaborated with mental health experts to generate a clinically sound labeling scheme and an annotated corpus to train machine learning and deep learning models.MethodsWe conducted a study using Bengali text-based data from blogs and open source platforms. We constructed a procedure for annotated corpus generation and extraction of textual information from Bengali literature for predictive analysis. We developed our own structured data set and designed a clinically sound labeling scheme with the help of mental health professionals, adhering to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) during the process. We used 5 machine learning models for detecting the severity of depression: kernel support vector machine (SVM), random forest, logistic regression K-nearest neighbor (KNN), and complement naive Bayes (NB). For the deep learning approach, we used long short-term memory (LSTM) units and gated recurrent units (GRUs) coupled with convolutional blocks or self-attention layers. Finally, we aimed for enhanced outcomes by using state-of-the-art pretrained language models.ResultsThe independent recurrent neural network (RNN) models yielded the highest accuracies and weighted F1 scores. GRUs, in particular, produced 81% accuracy. The hybrid architectures could not surpass the RNNs in terms of performance. Kernel SVM with term frequency–inverse document frequency (TF-IDF) embeddings generated 78% accuracy on test data. We used validation and training loss curves to observe and report the performance of our architectures. Overall, the number of available data remained the limitation of our experiment.ConclusionsThe findings from our experimental setup indicate that machine learning and deep learning models are fairly capable of assessing the severity of mental health issues from texts. For the future, we suggest more research endeavors to increase the volume of Bengali text data, in particular, so that modern architectures reach improved generalization capability.

Training and intrinsic evaluation of lightweight word embeddings for the clinical domain in Spanish.

Resources for Natural Language Processing (NLP) are less numerous for languages different from English. In the clinical domain, where these resources are vital for obtaining new knowledge about human health and diseases, creating new resources for the Spanish language is imperative. One of the most common approaches in NLP is word embeddings, which are dense vector representations of a word, considering the word's context. This vector representation is usually the first step in various NLP tasks, such as text classification or information extraction. Therefore, in order to enrich Spanish language NLP tools, we built a Spanish clinical corpus from waiting list diagnostic suspicions, a biomedical corpus from medical journals, and term sequences sampled from the Unified Medical Language System (UMLS). These three corpora can be used to compute word embeddings models from scratch using Word2vec and fastText algorithms. Furthermore, to validate the quality of the calculated embeddings, we adapted several evaluation datasets in English, including some tests that have not been used in Spanish to the best of our knowledge. These translations were validated by two bilingual clinicians following an ad hoc validation standard for the translation. Even though contextualized word embeddings nowadays receive enormous attention, their calculation and deployment require specialized hardware and giant training corpora. Our static embeddings can be used in clinical applications with limited computational resources. The validation of the intrinsic test we present here can help groups working on static and contextualized word embeddings. We are releasing the training corpus and the embeddings within this publication1.

Forest Tree Species Classification Based on Sentinel-2 Images and Auxiliary Data

Most research on forest tree species classification based on optical image data uses information such as spectral reflectance, vegetation index, texture, and phenology data. However, owing to the limited spectral resolution of multispectral images and the high cost of hyperspectral data, there is room for improvement in the classification of tree species in large areas based on optical images. The combined application of multispectral images and other auxiliary data can provide a new method for improving tree species classification accuracy. Hence, Sentinel-2 images were used to extract spectral reflectance, spectral index, texture, and phenological information. Data for topography, precipitation, air temperature, ultraviolet aerosol index, NO2 concentration, and other variables were included as auxiliary data. Models for forest tree species classification were constructed through feature combination and feature optimization using the random forest (RF), gradient tree boost (GTB), support vector machine (SVM), and classification and regression tree (CART) algorithms. The classification results of 16 feature combinations with the 4 classification methods were compared, and the contributions of different features to the classification models of forest tree species were evaluated. Finally, the optimal classification model was selected to identify the spatial distribution of forest tree species in the study area. The model based on feature optimization gave the best results among the 16 feature combination models. The overall accuracy and kappa coefficient were increased by 18% and 0.21, respectively, compared with the spectral classification model, and by 17% and 0.20, respectively, compared with the spectral and spectral index classification model. By analyzing the feature optimization model, it was found that terrain, ultraviolet aerosol index, and phenological information ranked as the top three features in terms of importance. Although the importance of spectral reflectance and spectral index features was lower, the number of feature variables accounted for a large proportion of the total. The importance of commonly used texture features was limited, and these features were not present in the feature optimization model. The RF algorithm had the highest classification accuracy, with an overall accuracy of 82.69% and a kappa coefficient of 0.80, among the four classification algorithms. The results of GTB were close to those of RF, and the difference in overall classification accuracy was only 0.14%. However, the results of the SVM and CART algorithms were relatively weaker, with overall classification accuracies of about 70%. It can be concluded that the combined application of Sentinel-2 images and auxiliary data can improve forest tree species classification accuracy. The model based on feature optimization achieved the highest classification accuracy among the 16 feature combination models. The spectral reflectance and spectral index data extracted from optical images are useful for tree species classification, but the effect of texture features was very limited. Auxiliary data, such as topographic features, ultraviolet aerosol index, phenological features, NO2 concentration features, topographic diversity features, precipitation features, temperature features, and multi-scale topographic location index data, can effectively improve forest tree species classification accuracy. The RF algorithm had the highest accuracy, and it can be used for tree species classification space distribution identification. The combined application of Sentinel-2 images and auxiliary data can improve classification accuracy, but the highest accuracy of the model was only 82.69%, which leaves room for improvement. Thus, more effective auxiliary data and the vertical structural parameters extracted from satellite LiDAR can be combined with multispectral images to improve forest tree species classification accuracy in future research.

TBR-NER: Research on COVID-19 Text Information Extraction Based on Joint Learning of Topic Recognition and Named Entity Recognition

There is a centralization of the core content in the text information of the new crown epidemic notification. This paper proposes a joint learning text information extraction method: TBR-NER (topic-based recognition named entity recognition) based on topic recognition and named entity recognition to predict the labeled risk areas and epidemic trajectory information in text information. Transfer learning and data augmentation are used to solve the problem of data scarcity caused by the initial local outbreak of the epidemic, and mutual understanding is achieved by topic self-labeling without introducing additional labeled data. Taking the epidemic cases in Hebei and Jilin provinces as examples, the reliability and effectiveness of the method are verified by five types of topic recognition and 15 types of entity information extraction. The experimental results show that, compared with the four existing NER methods, this method can achieve optimality faster through the mutual learning of each task at the early stage of training. The optimal accuracy in the independent test set can be improved by more than 20%, and the minimum loss value is significantly reduced. This also proves that the joint learning algorithm (TBR-NER) mentioned in this paper performs better in such tasks. The TBR-NER model has specific sociality and applicability and can help in epidemic prediction, prevention, and control.

Investigating Attention Mechanism for Page Object Detection in Document Images

Page object detection in scanned document images is a complex task due to varying document layouts and diverse page objects. In the past, traditional methods such as Optical Character Recognition (OCR)-based techniques have been employed to extract textual information. However, these methods fail to comprehend complex page objects such as tables and figures. This paper addresses the localization problem and classification of graphical objects that visually summarize vital information in documents. Furthermore, this work examines the benefit of incorporating attention mechanisms in different object detection networks to perform page object detection on scanned document images. The model is designed with a Pytorch-based framework called Detectron2. The proposed pipelines can be optimized end-to-end and exhaustively evaluated on publicly available datasets such as DocBank, PublayNet, and IIIT-AR-13K. The achieved results reflect the effectiveness of incorporating the attention mechanism for page object detection in documents.

A research on the effect of geostatistical texture analysis on image classification accuracy

In order to increase the accuracy results in various remote sensing applications, some additional parameters of texture information, which can provide information about the spatial relationship of a pixel in the image with other pixels, must be obtained. In addition to standard texture information extraction approaches, there is a need for approaches that take into account the spatial relationships of pixels in a certain neighborhood. In this study, it is aimed to reveal to what extent the use of characteristic information that can determine the texture properties of the image will affect the performance of the classification of satellite images by using the spatial dependency information obtained by geostatistical methods in satellite images. The data sets were classified with the Support Vector Machines (SVM) method, and according to pixel-based classification, an increase of 10% in agricultural area data and a 3% increase in urban area image was observed.

Enriching geometric digital twins of buildings with small objects by fusing laser scanning and AI-based image recognition

This paper addresses the challenge of enriching geometric digital twins of buildings, with a particular emphasis on capturing small but important entities from the electrical and the fire-safety domain, such as signs, sockets, switches, smoke alarms, etc. Unlike most previous research that focussed on structural elements and processed laser point clouds and images separately, we propose a novel method that fuses laser scanning and photogrammetry methods to capture the relevant objects, recognise them in 2D images and then map these to a 3D space. The considered object classes include electrical elements (light switch, light, speaker, socket, elevator button), safety elements (emergency switch, smoke alarm, fire extinguisher, escape sign), plumbing system elements (pipes), and other objects with useful information (door sign, board). Semantic information like class labels is extracted by applying AI-based image segmentation and then mapped to the 3D point cloud, segmenting the point cloud into point clusters. We subsequently fit geometric primitives to the point clusters and extract text information by AI-based text detection and recognition. The final output of our proposed method is an information-rich digital twin of buildings that contains geometric information, semantic information such as object categories and useful text information which is valuable in many aspects, like condition monitoring, facility maintenance and management. In summary, the paper presents a nearly fully-automated pipeline to enrich a geometric digital twin of buildings with details and provides a comprehensive case study.

Analysis on Hybrid Deep Neural Networks for Legal Domain Multitasks

An extensive quantity of online statistics accessible in the legal domain has made legal data processing the main sector of research development. A broad variety of problems, including legal document categorization, information extraction, and prediction have been put into a scope of legitimate system issues. The utilization of digitalized based inventive support has multi-fold advantages for the legal counsel community. These advantages comprise decreasing the laborious human task complicated in observant, extracting the relevant information, reducing the charge and time by-way-of automation, solving problems without the participation of law court otherwise with smaller period and attempt, arbitrating the constitution law for law professionals as well everyday users and building recommendations found on predictive analysis, which possibly examined additional perfect. In this chapter, we are analyzing the adaptation of various deep learning methods in the legal domain focusing on three main tasks namely text classification, information extraction, and prediction.

Texture Analysis of Enhanced MRI and Pathological Slides Predicts EGFR Mutation Status in Breast Cancer.

Objective Image texture information was extracted from enhanced magnetic resonance imaging (MRI) and pathological hematoxylin and eosin- (HE-) stained images of female breast cancer patients. We established models individually, and then, we combine the two kinds of data to establish model. Through this method, we verified whether sufficient information could be obtained from enhanced MRI and pathological slides to assist in the determination of epidermal growth factor receptor (EGFR) mutation status in patients. Methods We obtained enhanced MRI data from patients with breast cancer before treatment and selected diffusion-weighted imaging (DWI), T1 fast-spin echo (T1 FSE), and T2 fast-spin echo (T2 FSE) as the data sources for extracting texture information. Imaging physicians manually outlined the 3D regions of interest (ROIs) and extracted texture features according to the gray level cooccurrence matrix (GLCM) of the images. For the HE staining images of the patients, we adopted a specific normalization algorithm to simulate the images dyed with only hematoxylin or eosin and extracted textures. We extracted texture features to predict the expression of EGFR. After evaluating the predictive power of each model, the models from the two data sources were combined for remodeling. Results For enhanced MRI data, the modeling of texture information of T1 FSE had a good predictive effect for EGFR mutation status. For pathological images, eosin-stained images can achieve a better prediction effect. We selected these two classifiers as the weak classifiers of the final model and obtained good results (training group: AUC, 0.983; 95% CI, 0.95-1.00; accuracy, 0.962; specificity, 0.936; and sensitivity, 0.979; test group: AUC, 0.983; 95% CI, 0.94-1.00; accuracy, 0.943; specificity, 1.00; and sensitivity, 0.905). Conclusion The EGFR mutation status of patients with breast cancer can be well predicted based on enhanced MRI data and pathological data. This helps hospitals that do not test the EGFR mutation status of patients with breast cancer. The technology gives clinicians more information about breast cancer, which helps them make accurate diagnoses and select suitable treatments.