Extract Structural Information Research Articles

Structured information extraction from scientific text with large language models

Extracting structured knowledge from scientific text remains a challenging task for machine learning models. Here, we present a simple approach to joint named entity recognition and relation extraction and demonstrate how pretrained large language models (GPT-3, Llama-2) can be fine-tuned to extract useful records of complex scientific knowledge. We test three representative tasks in materials chemistry: linking dopants and host materials, cataloging metal-organic frameworks, and general composition/phase/morphology/application information extraction. Records are extracted from single sentences or entire paragraphs, and the output can be returned as simple English sentences or a more structured format such as a list of JSON objects. This approach represents a simple, accessible, and highly flexible route to obtaining large databases of structured specialized scientific knowledge extracted from research papers.

Using Abundant 1H Polarization to Enhance the Sensitivity of Solid-State NMR Spectroscopy.

Solid-state NMR spectroscopy has been playing a significant role in elucidating the structures and dynamics of materials and proteins at the atomic level for decades. As an extremely abundant nucleus with a very high gyromagnetic ratio, protons are widely present in most organic/inorganic materials. Thus, this Perspective highlights the advantages of proton detection at fast magic-angle spinning (MAS) and presents strategies to utilize and exhaust 1H polarization to achieve signal sensitivity enhancement of solid-state NMR spectroscopy, enabling substantial time savings and extraction of more structural and dynamics information per unit time. Those strategies include developing sensitivity-enhanced single-channel 1H multidimensional NMR spectroscopy, implementing multiple polarization transfer steps in each scan to enhance low-γ nuclei signals, and making full use of 1H polarization to obtain homonuclear and heteronuclear chemical shift correlation spectra in a single experiment. Finally, outlooks and perspectives are provided regarding the challenges and future for the further development of sensitivity-enhanced proton-based solid-state NMR spectroscopy.

Parsing of Research Documents into XML Using Formal Grammars

Automatic information extraction of content and style format in paged documents is challenging. It requires the conversion of the original document into a granular level of details for which every document section and content is identifiable. This functionality or tool does not exist for any academic research document yet. In this paper, we present an automated process of parsing research paper documents into XML files using a formal method approach of context-free grammars (CFGs) and regular expressions (REGEXs) definable of a standard template. We created a tool for the algorithms to parse these documents into tree-like structures organized as XML files named research_XML (RX) parser. The RX tool performed the extraction of syntactic structure and semantic information of the document’s contents into XML files. These XML output files are lightweight, analyzable, query-able, and web interoperable. The RX tool has a success rate of 91% when evaluated on fifty varying research documents of 160 average pages and 8,004 total pages. The tool and test data are accessible on GitHub repo. The novelty of our process is specific to applying formal techniques for information extraction in structured multipaged documents and academic research documents thus advancing the research in automatic information extraction.

Seismic Thermography

ABSTRACT Variations in temperature within the Earth are of great interest because they indicate the thickness and, consequently, mechanical strength of the lithosphere and density variations and convection patterns in the sublithospheric mantle. Seismic tomography maps seismic velocity variations in the mantle, which strongly depend on temperature. Temperatures are, thus, often inferred from tomography. Tomographic models, however, are nonunique solutions of inverse problems, regularized to ensure model smoothness or small model norm, not plausible temperature distributions. For example, lithospheric geotherms computed from seismic velocity models typically display unrealistic oscillations, with improbable temperature decreases with depth within shallow mantle lithosphere. The errors due to the intermediate-model nonuniqueness are avoided if seismic data are inverted directly for temperature. The recently developed thermodynamic inversion methods use computational petrology and thermodynamic databases to jointly invert seismic and other data for temperature and composition. Because seismic velocity sensitivity to composition is much weaker than to temperature, we can invert seismic data primarily for temperature, with reasonable assumptions on composition and other relevant properties and with additional inversion parameters such as anisotropy. Here, we illustrate thus-defined seismic thermography with thermal imaging of the lithosphere and asthenosphere using surface waves. We show that the accuracy of the models depends critically on the accuracy of the extraction of structural information from the seismic data. Random errors have little effect but correlated errors of even a small portion of 1% can affect the models strongly. We invert data with different noise characteristics and test a simple method to estimate phase velocity errors. Seismic thermography builds on the techniques of seismic tomography and relies on computational petrology, but it is emerging as a field with its scope of goals, technical challenges, and methods. It produces increasingly accurate models of the Earth, with important inferences on its dynamics and evolution.

Experimental Study on Measuring and Tracking Structural Displacement Based on Surveillance Video Image Analysis.

The digital image method of monitoring structural displacement is receiving more attention today, especially in non-contact structure health monitoring. Some obvious advantages of this method, such as economy and convenience, were shown while it was used to monitor the deformation of the bridge structure during the service period. The image processing technology was used to extract structural deformation feature information from surveillance video images containing structural displacement in order to realize a new non-contact online monitoring method in this paper. The influence of different imaging distances and angles on the conversion coefficient (η) that converts the pixel coordinates to the actual displacement was first studied experimentally. Then, the measuring and tracking of bridge structural displacement based on surveillance video images was investigated by laboratory-scale experiments under idealized conditions. The results showed that the video imaging accuracy can be affected by changes in the relative position of the imaging device and measured structure, which is embodied in the change in η (actual size of individual pixel) on the structured image. The increase in distance between the measured structure and the monitoring equipment will have a significant effect on the change in the η value. The value of η varies linearly with the change in shooting distance. The value of η will be affected by the changes in shooting angle. The millimeter-level online monitoring of the structure displacement can be realized using images based on surveillance video images. The feasibility of measuring and tracking structural displacement based on surveillance video images was confirmed by a laboratory-scale experiment.

Concrete crack segmentation based on multi-dimensional structure information fusion-based network

The state-of-the-art (SOTA) crack segmentation methods still face the challenges in low-contrast shallow crack recognition and multi-scene compatibility. Therefore, this paper proposes a solution based on a multi-dimensional structure information fusion-based network (MDSIFNet). This network consists of two branches, one for extracting two-dimensional (2D) spatial feature information and another for acquiring three-dimensional (3D) geometric one. In the former, the designed 2D curve structure constraint module based on prior knowledge combined with a pretrained 2D feature extraction module can reduce learning samples and realize crack segmentation in unseen scenes. In the latter, a one-hot transformation and a 3D geometric structure information extraction module are designed to make the network pay attention to the geometric features of shallow cracks and improve their segmentation accuracy. Finally, a fusion module couples the multi-dimensional structure feature information of these two branches and outputs pixel-wise segmentation results that can be used for crack measurement and quantitative analysis. The experimental results show that the designed model network MDSIFNet performs better than the SOTA methods in terms of performance and visualized results.

Deep convolutional neural networks to restore single-shot electron microscopy images

Advanced electron microscopy techniques, including scanning electron microscopes (SEM), scanning transmission electron microscopes (STEM), and transmission electron microscopes (TEM), have revolutionized imaging capabilities. However, achieving high-quality experimental images remains a challenge due to various distortions stemming from the instrumentation and external factors. These distortions, introduced at different stages of imaging, hinder the extraction of reliable quantitative insights. In this paper, we will discuss the main sources of distortion in TEM and S(T)EM images, develop models to describe them, and propose a method to correct these distortions using a convolutional neural network. We validate the effectiveness of our method on a range of simulated and experimental images, demonstrating its ability to significantly enhance the signal-to-noise ratio. This improvement leads to a more reliable extraction of quantitative structural information from the images. In summary, our findings offer a robust framework to enhance the quality of electron microscopy images, which in turn supports progress in structural analysis and quantification in materials science and biology.

DiffMAR: A Generalized Diffusion Model for Metal Artifact Reduction in CT images.

X-ray imaging frequently introduces varying degrees of metal artifacts to computed tomography (CT) images when metal implants are present. For the metal artifact reduction (MAR) task, existing end-to-end methods often exhibit limited generalization capabilities. While methods based on multiple iterations often suffer from accumulative error, resulting in lower-quality restoration outcomes. In this work, we innovatively present a generalized diffusion model for Metal Artifact Reduction (DiffMAR). The proposed method utilizes a linear degradation process to simulate the physical phenomenon of metal artifact formation in CT images and directly learn an iterative restoration process from paired CT images in the reverse process. During the reverse process of DiffMAR, a Time-Latent Adjustment (TLA) module is designed to adjust time embedding at the latent level, thereby minimizing the accumulative error during iterative restoration. We also designed a structure information extraction (SIE) module to utilize linear interpolation data in the image domain, guiding the generation of anatomical structures during the iterative restoring. This leads to more accurate and robust shadow-free image generation. Comprehensive analysis, including both synthesized data and clinical evidence, confirms that our proposed method surpasses the current state-of-the-art (SOTA) MAR methods in terms of both image generation quality and generalization.

Structural descriptors and information extraction from X-ray emission spectra: aqueous sulfuric acid.

Machine learning can reveal new insights into X-ray spectroscopy of liquids when the local atomistic environment is presented to the model in a suitable way. Many unique structural descriptor families have been developed for this purpose. We benchmark the performance of six different descriptor families using a computational data set of 24 200 sulfur Kβ X-ray emission spectra of aqueous sulfuric acid simulated at six different concentrations. We train a feed-forward neural network to predict the spectra from the corresponding descriptor vectors and find that the local many-body tensor representation, smooth overlap of atomic positions and atom-centered symmetry functions excel in this comparison. We found a similar hierarchy when applying the emulator-based component analysis to identify and separate the spectrally relevant structural characteristics from the irrelevant ones. In this case, the spectra were dominantly dependent on the concentration of the system, whereas adding the second most significant degree of freedom in the decomposition allowed for distinction of the protonation state of the acid molecule.

An End-to-End Tag Recognition Architecture for Industrial Meter

Optical character recognition (OCR) technology is promoting the process of automation in daily information recording and inspection of industrial meters. However, bad installation scenes and long-time running cause industrial meters to age to varying degrees, which increases the difficulty of OCR. Besides, in practice, operators typically use handheld cameras to extract tag information from industrial meters. During the shooting process, extreme lighting, free shooting angles, and shooting distance also cause many difficulties for OCR. Considering such difficulties, an end-to-end recognition architecture is developed to obtain a better OCR performance. The proposed architecture can quickly extract structured information from normal or skewed text images. A novel yolov5_adaloss with a specific penalty factor is designed to alleviate the influence of illumination, installation scene, age, skew, and distance on the classification accuracy of Tags. For images with large skew angles, a mathematical method is proposed to calculate the text inclination angle for better OCR performance. The contribution of this work is twofold. Firstly, the architecture proposed is lightweight, which only needs a low computing cost and a short inference time. Secondly, as a practical application-oriented architecture, this work does not require much training, labeling, and fine-tuning, which is easy to generalize to other structured text recognition tasks. Experiments show that the method proposed in this paper can achieve excellent performance in meter tag recognition tasks on actual industrial images of meters.

Low-Light Enhancement Method Based on a Retinex Model for Structure Preservation

Enhancing low-light image visibility is a critical task in computer vision since it helps to improve input for high-level algorithms. High-quality images typically have clear structural information. In previous studies, due to the lack of proper structural guidance, restored images had some problems, such as unclear structural areas and overexposed or underexposed local areas. To address the above problems, in this paper, we introduce a coefficient of variation (COV) with excellent performance in maintaining structural information, and then we propose a low-light image enhancement method that utilizes the COV to extract structural information from images. First, we apply a traditional retinex model to estimate both reflectance and illumination. Second, we use the COV to indicate the degree of dispersion of the input sample, which enables us to obtain a robust structure-distinguishing weight map for low-light images. The weight map is adaptively divided to obtain a structural weight map, which is then used to enhance the gradient image. This process is applied before the reflectance layer of the retinex model. Finally, the result is obtained by using the block coordinate descent method. According to extensive experiments, outstanding results can be achieved by our proposed method in terms of both subjective and objective evaluation metrics in comparison with other state-of-the-art methods. The source code is available at our website <xref ref-type="fn" rid="fn1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">¹</xref> <fn id="fn1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><label>¹</label> <uri>https://github.com/bbxavi/spcv22</uri> </fn> .

Perspectives on solution-based small angle X-ray scattering for protein and biological macromolecule structural biology.

Small-angle X-ray scattering (SAXS) is used to extract structural information from a wide variety of non-crystalline samples in different fields (e.g., materials science, physics, chemistry, and biology). This review provides an overview of SAXS as applied to structural biology, specifically for proteins and other biomacromolecules in solution with an emphasis on extracting key structural parameters and the interpretation of SAXS data using a diverse array of techniques. These techniques cover aspects of building and assessing models to describe data measured from monodispersed and ideal dilute samples through to more complicated structurally polydisperse systems. Ab initio modelling, rigid body modelling as well as normal-mode analysis, molecular dynamics, mixed component and structural ensemble modelling are discussed. Dealing with polydispersity both physically in terms of component separation as well as approaching the analysis and modelling of data of mixtures and evolving systems are described, including methods for data decomposition such as single value decomposition/principle component analysis and evolving factor analysis. This review aims to highlight that solution SAXS, with the cohort of developments in data analysis and modelling, is well positioned to build upon the traditional 'single particle view' foundation of structural biology to take the field into new areas for interpreting the structures of proteins and biomacromolecules as population-states and dynamic structural systems.

Development of online application for crime reporting and handling in Malawi police service (chatbot)

A chatbot is a conversational device designed to provide smart communication with humans through speech and text. Chatbots emerge as virtual machines that interact with humans through their dialect, offering both spoken and textual conversation and responses to questions. In recent times, there has been tremendous progress in the development of virtual assistants and chatbots with advanced technologies based on speech recognition. This paper presents a chatbot for the web-based Crime Reporting and Handling System in the Malawi Police Service. The chatbot utilizes a combination of classification and generative models. It features a complaint registration system that allows users to file complaints. A custom named entity recognition model is employed to extract structured information such as location, time, and crime type from unstructured complaints, enabling authorities to comprehend complaints effectively and efficiently. The chatbot strives to provide efficient and user-friendly methods for registering complaints using these techniques, as well as informing individuals about the crime reporting and handling system. Thus, the system aspires to apply Natural Language Processing (NLP) for social good. In this paper, we will discuss how chatbots are designed, the machine learning and deep learning techniques and algorithms used, compare the technologies, and evaluate the performance of chatbots. Chatbots are becoming friendlier to humans, evolving beyond mere communication machines and representing future aspects of human and technology interaction.

Path Updating of the Legal Ontology Model for Personal Information Protection in the Age of Big Data

Abstract This paper proposes a legal ontology model for personal information protection to extract structured information on the high-frequency elements of common privacy leakage and data security cases as key information of judicial documents. The prediction model of relevant sentences based on the CART decision tree is used to predict the outcome of the defendant’s sentence. The application scenarios of the legal ontology model of personal information protection in judicial guarantee are explored to guarantee the fairness of adjudication of personal information protection cases in terms of reliability of legal counseling, accuracy of sentencing assistance, and categorization of cases. The results show that 113 sample cases were retrieved by keyword search, excluding duplicate cases, with Beijing having the highest number of 42. With the development of the big data era, there has been a significant increase in personal information protection cases after 2019. The legal ontology model constructed in this paper has only 56 instances of classification error, the Macro-F1-Score and Micro-F1-Score indicator values are above 70%, and the utility values of personal information protection-related cases are all higher than 0.5. Therefore, the legal ontology model for personal information protection constructed in this paper is accurate in classification and sentencing prediction, with high reliability, and realizes the path of the big data era.

Low-tubal-rank tensor completion via local and nonlocal knowledge

Many previous studies have indicated that exploiting prior information about the underlying tensor data is a sound approach for completing missing or damaged entries. This prior information can be divided into three commonly used categories: low-rankness, local piecewise smoothness, and nonlocal self-similarity (NSS) priors. Although existing methods based on these priors have gained considerable attention, the majority of studies utilize only one or two of these priors, leading to inadequate extraction of structural information from tensors. To address this limitation and comprehensively depict the inherent structural information of underlying tensor data, this article develops a novel tensor completion framework that can simultaneously utilize the three abovementioned priors within a plug-and-play framework. More precisely, we adopt the tensor correlated total variation (t-CTV) norm as a robust representation for capturing the combined effects of low-rankness and local piecewise smoothness priors, eliminating the need for a trade-off parameter in the process; meanwhile, we introduce an advanced denoiser to explore the NSS prior. Furthermore, to address the presented optimization model, we design an alternating direction method of multipliers (ADMM) algorithm and innovatively provide its corresponding global convergence guarantees. Extensive numerical experiments on real tensor data, including color, medical and hyperspectral images, demonstrate that the proposed method surpasses various advanced approaches in terms of both quality metrics and visual effects.

Effortless Extraction of Structural and Orientational Information from 13C-1H Dipolar Couplings for Thiophene Mesogens.

Five molecular mesogens containing phenyl rings and thiophene are subjected to a detailed 13C NMR investigation. The first mesogen contains only phenyl rings, while the other four have thiophene with substitution at position 2 or 3. Two of these also have a spacer inserted between the thiophene and the rest of the core unit. The mesophase properties evaluated by complementary techniques reveal an enantiotropic nematic phase for all the cases and smectic C as well as Crystal J for a few mesogens. In addition to solution 13C NMR, the samples were studied using 1D and 2D solid-state 13C NMR experiments in the liquid crystalline phase. The chemical shifts and 13C-1H dipolar couplings obtained in the mesophase provided cutting-edge information about the molecular structure and orientation of the thiophene mesogens. Accordingly, dramatic differences in these parameters are noted for the mesogens, and consequently, the identification of 2- and 3-substituted thiophene mesogens is accomplished by a simple visual examination of the 2D spectra. Furthermore, for mesogens with a spacer between thiophene and the rest of the core, 13C chemical shifts and 13C-1H dipolar couplings showed remarkable variation, which was directly reflected in the order parameters. For instance, the order parameter (Szz) of thiophene in 2- and 3-substituted mesogens in which the spacer is absent is ∼0.63 whereas for those with spacer, it is reduced to ∼0.14-0.18. In comparison, the mesogen in which the core unit is made up of phenyl rings alone that is used to benchmark the characteristics of thiophene ordering showed an order parameter of ∼0.85. The study unambiguously demonstrates the supremacy of 13C NMR in extracting the structural and orientational information on mesogens in which thiophene is a constituent of the core unit.

Generating psychological analysis tables for children's drawings using deep learning

The usefulness of drawing-based psychological testing has been demonstrated in a variety of studies. By using the familiar medium of drawing, drawing-based psychological testing can be applied to a wide range of age groups and is particularly effective with children who have difficulty expressing themselves verbally. Drawing tests are usually implemented face-to-face, requiring specialized counseling staff, and can be time-consuming and expensive to apply to large numbers of children. These problems seem to be solved by applying highly developed artificial intelligence techniques. If artificial intelligence (AI) can analyze children's drawings and perform psychological analysis, it will be possible to use it as a service and take tests online or through smartphones. There have been various attempts to automate the drawing of psychological tests by utilizing deep learning technology to process images. Previous studies using classification have been limited in their ability to extract structural information. In this paper, we analyze the House-Tree-Person Test (HTP), one of the drawing psychological tests widely used in clinical practice, by utilizing object detection technology that can extract more diverse information from images. In addition, we extend the existing research that has been limited to the extraction of relatively simple psychological features and generate a psychological analysis table based on the extracted features that can be used to assist experts in the process of psychological testing. Our research findings indicate that the object detection performance achieves a mean Average Precision (mAP) of approximately 92.6∼94.1 %, and the average accuracy of the psychological analysis table is 94.4 %.

Pseq2Sites: Enhancing protein sequence-based ligand binding-site prediction accuracy via the deep convolutional network and attention mechanism

Protein-ligand interactions play an essential role in many biological processes, and prior knowledge of ligand binding sites is necessary for successful drug design. Many 3D structure- and sequence-based methods have been proposed for identifying ligand binding sites. The 3D structure-based methods typically achieve better binding site prediction than the sequence-based methods. However, as deep-learning techniques that can extract structural information from large-scale sequence data have been developed, the performance gap between 3D structure- and sequence-based methods is narrowing. Nonetheless, there remains room for improvement in sequence-based prediction. We propose Pseq2Sites, a sequence-based deep-learning model for predicting ligand binding sites. Pseq2Sites comprises a 1D convolutional neural network that extracts local features from the protein sequence, and a position-based attention mechanism that captures long-distance dependencies between binding residues. To verify the effectiveness of the proposed method, we compared it with other state-of-the-art methods using three public datasets: COACH420, HOLO4K, and CSAR-NRC HiQ. Utilizing solely protein sequence information, Pseq2Sites outperformed 3D structure-based state-of-the-art methods on external test datasets; within the COACH420 dataset, Pseq2Sites remarkably identified 97% of the binding pockets (at a significance level δ = 0.5), which was 27% higher than the second highest-performing model. Pseq2Sites also achieved outstanding binding site prediction, even for proteins with low similarity to the training dataset. Our code is available at https://github.com/Blue1993/Pseq2Sites.

User's interface for extraction of the chemical structure information from the systematic name of organic compound

The user's interface «Nomenclature Generator» for extraction of the chemical structure information from the systematic name of organic compound represented according to IUPAC nomenclature is developed at the All-Russian Institute for Scientific and Technical Information of Russian Academy of Sciences.

Extraction of Interoperable Data from Healthcare Documents by Identifying Common Data Elements: An Analysis of Radiation Therapy Planning CT Physician Order Entry Records

Introduction: Documentation as well as IT-based management of medical data is of ever-increasing relevance in modern medicine. As radiation oncology is a rather technical, data-driven discipline, standardization, and data exchange are in principle possible. We examined electronic healthcare documents to extract structured information. Planning CT order entry documents were chosen for the analysis, as this covers a common and structured step in radiation oncology, for which standardized documentation may be achieved. The aim was to examine the extent to which relevant information may be exchanged among different institutions. Materials and Methods: We contacted representatives of nine radiation oncology departments. Departments using standardized electronic documentation for planning CT were asked to provide templates of their records, which were analyzed in terms of form and content. Structured information was extracted by identifying definite common data elements, containing explicit information. Relevant common data elements were identified and classified. A quantitative analysis was performed to evaluate the possibility of data exchange. Results: We received data of seven documents that were heterogeneous regarding form and content. 181 definite common data elements considered relevant for the planning CT were identified and assorted into five semantic groups. 139 data elements (76.8%) were present in only one document. The other 42 data elements were present in two to six documents, while none was shared among all seven documents. Conclusion: Structured and interoperable documentation of medical information can be achieved using common data elements. Our analysis showed that a lot of information recorded with healthcare documents can be presented with this approach. Yet, in the analyzed cohort of planning CT order entries, only a few common data elements were shared among the majority of documents. A common vocabulary and consensus upon relevant information is required to promote interoperability and standardization.