Unstructured Textual Information Research Articles

AI-Driven Diagnostic Assistance in Medical Inquiry: Reinforcement Learning Algorithm Development and Validation.

For medical diagnosis, clinicians typically begin with a patient's chief concerns, followed by questions about symptoms and medical history, physical examinations, and requests for necessary auxiliary examinations to gather comprehensive medical information. This complex medical investigation process has yet to be modeled by existing artificial intelligence (AI) methodologies. The aim of this study was to develop an AI-driven medical inquiry assistant for clinical diagnosis that provides inquiry recommendations by simulating clinicians' medical investigating logic via reinforcement learning. We compiled multicenter, deidentified outpatient electronic health records from 76 hospitals in Shenzhen, China, spanning the period from July to November 2021. These records consisted of both unstructured textual information and structured laboratory test results. We first performed feature extraction and standardization using natural language processing techniques and then used a reinforcement learning actor-critic framework to explore the rational and effective inquiry logic. To align the inquiry process with actual clinical practice, we segmented the inquiry into 4 stages: inquiring about symptoms and medical history, conducting physical examinations, requesting auxiliary examinations, and terminating the inquiry with a diagnosis. External validation was conducted to validate the inquiry logic of the AI model. This study focused on 2 retrospective inquiry-and-diagnosis tasks in the emergency and pediatrics departments. The emergency departments provided records of 339,020 consultations including mainly children (median age 5.2, IQR 2.6-26.1 years) with various types of upper respiratory tract infections (250,638/339,020, 73.93%). The pediatrics department provided records of 561,659 consultations, mainly of children (median age 3.8, IQR 2.0-5.7 years) with various types of upper respiratory tract infections (498,408/561,659, 88.73%). When conducting its own inquiries in both scenarios, the AI model demonstrated high diagnostic performance, with areas under the receiver operating characteristic curve of 0.955 (95% CI 0.953-0.956) and 0.943 (95% CI 0.941-0.944), respectively. When the AI model was used in a simulated collaboration with physicians, it notably reduced the average number of physicians' inquiries to 46% (6.037/13.26; 95% CI 6.009-6.064) and 43% (6.245/14.364; 95% CI 6.225-6.269) while achieving areas under the receiver operating characteristic curve of 0.972 (95% CI 0.970-0.973) and 0.968 (95% CI 0.967-0.969) in the scenarios. External validation revealed a normalized Kendall τ distance of 0.323 (95% CI 0.301-0.346), indicating the inquiry consistency of the AI model with physicians. This retrospective analysis of predominantly respiratory pediatric presentations in emergency and pediatrics departments demonstrated that an AI-driven diagnostic assistant had high diagnostic performance both in stand-alone use and in simulated collaboration with clinicians. Its investigation process was found to be consistent with the clinicians' medical investigation logic. These findings highlight the diagnostic assistant's promise in assisting the decision-making processes of health care professionals.

Semantic Mapping of Named-Entities in openEHR Templates and Ad-hoc Generation of Compositions.

Integration of free texts from reports written by physicians to an interoperable standard is important for improving patient-centric care and research in the medical domain. In the context of unstructured clinical data, NLP Information Extraction serves in finding information in unstructured text. To our best knowledge, there is no efficient solution, in which extracted Named-Entities of an NLP pipeline can be ad-hoc inserted in openEHR compositions. We therefore developed a software solution that solves this data integration problem by mapping Named-Entities of an NLP pipeline to the fields of an openEHR template. The mapping can be accomplished by any user without any programming intervention and allows the ad-hoc creation of a composition based on the mappings.

SPECE: Subject Position Encoder in Complex Embedding for Relation Extraction

As a crucial component of many natural language processing tasks, extracting entities and relations transforms unstructured text information into structured data, providing essential support for constructing knowledge graphs (KGs). However, current entity relation extraction models often prioritize the extraction of richer semantic features or the optimization of relation extraction methods, overlooking the significance of positional information and subject characteristics in this task. To solve this problem, we introduce the subject position-based complex exponential embedding for entity relation extraction model (SPECE). The encoder module of this model ingeniously combines a randomly initialized dilated convolutional network with a BERT encoder. Notably, it determines the initial position of the predicted subject based on semantic cues. Furthermore, it achieves a harmonious integration of positional encoding features and textual features through the adoption of the complex exponential embedding method. The experimental outcomes on both the NYT and WebNLG datasets reveal that, when compared to other baseline models, our proposed SPECE model demonstrates significant improvements in the F1 score on both datasets. This further validates its efficacy and superiority.

Computer-Assisted Analysis of Qualitative News Dissemination

This study delves into the potential of utilizing deep learning (DL) techniques to analyze qualitative news dissemination for trading purposes. DL, renowned for its prowess in handling vast datasets and deciphering intricate patterns, holds promise in aiding investors seeking to enhance their trading strategies. Specifically, Long Short-Term Memory (LSTM) networks, known for their capacity to retain contextual information, are explored in this research. By employing DL models, we aim to forecast market sentiment based on news headlines, focusing on the Dow Jones industrial average from 2008 to 2020. Leveraging 25 daily news headlines, we extend our analysis to develop an algorithmic trading strategy. Through rigorous testing across two distinct cases over five-time steps, our study evaluates the effectiveness of DL-driven approaches in real-world trading scenarios. Furthermore, this research contributes to the growing body of literature on the intersection of deep learning and financial markets. By examining the application of DL in qualitative news analysis for trading purposes, we provide insights into the potential implications for investors and financial analysts. The findings offer valuable guidance for leveraging advanced computational techniques for decision-making in dynamic market environments. Additionally, this study underscores the importance of incorporating qualitative news data into trading strategies, highlighting the role of DL in extracting meaningful signals from unstructured textual information’s Overall, our findings shed light on the opportunities and challenges associated with harnessing DL for trading on news sentiment.

Demand Planning: Riding Disruptive Wave of AI and Accelerated Computing

Traditional demand planning methods often struggle to keep pace with the complexity, volatility, and vast datasets inherent in modern supply chains. Artificial intelligence (AI) offers a transformative solution, revolutionizing demand planning with its ability to analyze vast amount of data, identify complex patterns, and generate highly accurate forecasts. This paper explores the latest advancements in AI for demand planning, encompassing machine learning, deep learning, and natural language processing (NLP). The focus is on how these techniques enhance demand sensing capabilities, incorporating real-time market signals, external data sources, and unstructured text information. Furthermore, the potential of AI to optimize inventory management, enable scenario planning, and increase supply chain resilience in response to unexpected disruptions are examined. The paper also addresses practical challenges in implementing AI-powered demand planning solutions, and outlines areas for future research. Most importantly, the paper provides the robust methologies to integrate the emerging AI developments in demand planning process.

Sensing complicated meanings from unstructured data: a novel hybrid approach

The majority of data on computers nowadays is in the form of unstructured data and unstructured text. The inherent ambiguity of natural language makes it incredibly difficult but also highly profitable to find hidden information or comprehend complex semantics in unstructured text. In this paper, we present the combination of natural language processing (NLP) and convolution neural network (CNN) hybrid architecture called automated analysis of unstructured text using machine learning (AAUT-ML) for the detection of complex semantics from unstructured data that enables different users to make understand formal semantic knowledge to be extracted from an unstructured text corpus. The AAUT-ML has been evaluated using three datasets data mining (DM), operating system (OS), and data base (DB), and compared with the existing models, i.e., YAKE, term frequency-inverse document frequency (TF-IDF) and text-R. The results show better outcomes in terms of precision, recall, and macro-averaged F1-score. This work presents a novel method for identifying complex semantics using unstructured data.

Pediatric injuries classification using openAI: new opportunities for epidemiological surveillance

Abstract Background The burden of unintentional pediatric injuries remains a significant concern for healthcare systems worldwide. Timely and accurate surveillance is critical for injury prevention and resource allocation. However, the diagnosis reported in the emergency department records is often not coded, and the manual coding of these records is labor-intensive, time-consuming, and prone to errors. To implement an automatic coding system using GPT-based models to extract and classify injury data in the Italian language from pediatric emergency department records. Methods The study included 283,468 admission records to the pediatric Emergency Department of Padova University Hospital from 2007 to 2018. Each access is mandatorily registered in an electronic data collection system. For each emergency department access, both coded (administrative and demographic data) and free-text (diagnosis) information are reported. A random subset of 40,030 records underwent classification of free-text diagnosis (as injury or not) by an expert clinician (gold standard). OpenAI was used for the classification task. OpenAI is an extensive language model based on the GPT architecture. Specifically, the GPT-3.5 variant was employed for the present work. openAI was accessed through a public application programming interface (API) using R software. Results Preliminary results showed a classification accuracy of 96.2%. The tool's ability to correctly classify the injuries (sensitivity) was 95%, while the specificity was 96.5%. Conclusions The performance of the classification task was excellent. The present results demonstrate the feasibility of GPT-based models for processing unstructured free text information from medical records. Key messages • OpenAI would represent a promising opportunity to improve pediatric injuries surveillance • OpenAI would improve the classification of unstructured information from medical records

Quantitative feature extraction of unstructured data from GitLab BioAI pathology reports of cancer using an enhanced RPA NLP method

Unstructured pathology report plays a major role in definitive cancer diagnosis. Accessing or searching unstructured textual information from the clinical pathology reports is one of the major concerns in cancer healthcare sector to provide precise medicine, analysis of cancer outcomes, providing cancer care services, accurate measurement for future prediction, treatment history, and comparative future research work. An efficient methodology has to be introduced for to extract quantitative information from the unstructured cancer data. Integrating computational intelligence in Robotic Process Automation can be done to process this data and automate repetitive activities for evaluating patients clinical pathology report. RPA-based NLP BERT system is designed and evaluated to automatically extract information on these variables for the patients from pathology report. In order to detect tumour and outcomes from documented pathology reports, a supervised machine learning keyword based extraction algorithm was developed in which the pathology data are examined to extract keywords from 2087 reports with 1579 of data reports being processed for the development phase and 508 of data being used for evaluation. The precision recall and accuracy are calculated for organ specimens for cancer test as (0.984, 0.982, 0.9839), test methodology(0.986, 0.981,0.9956) and pathological result(0.986, 0.9938, 0.9795) were achieved. The feasibility of autonomously extracting pre-defined data from clinical narratives for cancer research were established in this work. The outcomes showed that our methodology was suitable for actual use in obtaining essential information from pathology reports.

Asset pricing via fused deep learning with visual clues

Asset pricing via machine learning provides a promising way to capture price trends by fusing heterogeneous market factors to analyze their joint impact on stock movements rather than relying on statistical and econometric models in finance to explore the causality between a market indicator and stock returns. However, the fusion nature of machine learning also hides the way to unveil the internal mechanism of stock movements. In this study, a deep learning framework with visual clues is presented to unveil the entangled factors and their function on stock movements. In particular, a context-aware hierarchical attention mechanism (CHARM) is first proposed to encode unstructured textual media information to trace the literal power of news on such media-aware stock movements. The encoded media and other structured market factors are further fused via tensor-based learning to infer and visualize their interactions on stock fluctuations. Last, a pre-estimating method for locating turning points as trading clues is utilized to improve the efficiency of each investment opportunity. Experiments conducted in real securities markets demonstrate the proposed framework not only improve the readability of the investing strategies, but also enhance the predictive accuracy and the investing returns comparing with the state-of-the-art models, including AZFinText, TeSIA, eLSTM, CMT, MAC and SA-DLSTM.

Acute stroke CDS: automatic retrieval of thrombolysis contraindications from unstructured clinical letters.

Thrombolysis treatment for acute ischaemic stroke can lead to better outcomes if administered early enough. However, contraindications exist which put the patient at greater risk of a bleed (e.g. recent major surgery, anticoagulant medication). Therefore, clinicians must check a patient's past medical history before proceeding with treatment. In this work we present a machine learning approach for accurate automatic detection of this information in unstructured text documents such as discharge letters or referral letters, to support the clinician in making a decision about whether to administer thrombolysis. We consulted local and national guidelines for thrombolysis eligibility, identifying 86 entities which are relevant to the thrombolysis decision. A total of 8,067 documents from 2,912 patients were manually annotated with these entities by medical students and clinicians. Using this data, we trained and validated several transformer-based named entity recognition (NER) models, focusing on transformer models which have been pre-trained on a biomedical corpus as these have shown most promise in the biomedical NER literature. Our best model was a PubMedBERT-based approach, which obtained a lenient micro/macro F1 score of 0.829/0.723. Ensembling 5 variants of this model gave a significant boost to precision, obtaining micro/macro F1 of 0.846/0.734 which approaches the human annotator performance of 0.847/0.839. We further propose numeric definitions for the concepts of name regularity (similarity of all spans which refer to an entity) and context regularity (similarity of all context surrounding mentions of an entity), using these to analyse the types of errors made by the system and finding that the name regularity of an entity is a stronger predictor of model performance than raw training set frequency. Overall, this work shows the potential of machine learning to provide clinical decision support (CDS) for the time-critical decision of thrombolysis administration in ischaemic stroke by quickly surfacing relevant information, leading to prompt treatment and hence to better patient outcomes.

WISK: A Workload-aware Learned Index for Spatial Keyword Queries

Spatial objects often come with textual information, such as Points of Interest (POIs) with their descriptions, which are referred to as geo-textual data. To retrieve such data, spatial keyword queries that take into account both spatial proximity and textual relevance have been extensively studied. Existing indexes designed for spatial keyword queries are mostly built based on the geo-textual data without considering the distribution of queries already received. However, previous studies have shown that utilizing the known query distribution can improve the index structure for future query processing. In this paper, we propose WISK, a learned index for spatial keyword queries, which self-adapts for optimizing querying costs given a query workload. One key challenge is how to utilize both structured spatial attributes and unstructured textual information during learning the index. We first divide the data objects into partitions, aiming to minimize the processing costs of the given query workload. We prove the NP-hardness of the partitioning problem and propose a machine learning model to find the optimal partitions. Then, to achieve more pruning power, we build a hierarchical structure based on the generated partitions in a bottom-up manner with a reinforcement learning-based approach. We conduct extensive experiments on real-world datasets and query workloads with various distributions, and the results show that WISK outperforms all competitors, achieving up to 8× speedup in querying time with comparable storage overhead.

DETECTION OF COMPLEX SEMANTICS FROM UNSTRUCTURED DATA USING NLP & CONVOLUTION NEURAL NETWORK TECHNIQUE

The majority of data on computers nowadays still takes the form of unstructured text. The inherent ambiguity of natural language makes it incredibly difficult but also highly profitable to find hidden information or comprehend complex semantics in unstructured text. In this paper, we present combination of Natural Language Processing (NLP) and Convolution Neural Network (CNN) hybrid architecture for detection of complex semantics from unstructured data that enables different users to make understand formal semantic knowledge to be extracted from an unstructured text corpus. Keywords— Convolution Neural Network, Natural Language Processing, Controlled Natural Language, Text Mining, Information Extraction.

A HYBRID QUANTUM-PERFECTED MODEL OF ARTIFICIAL INTELLIGENCE IN THE PROBLEM OF AUTOMATIC RECOGNITION AND FAST CONVERSION OF UNSTRUCTURED TEXT INFORMATION INTO SPATIAL

Background. Efficiently converting large amounts of unstructured text data into spatial information is crucial for managing water distribution systems. This allows for the conversion of extensive sets of text information, such as reports, orders, letters, and other documents, into point classes of spatial objects in geographic information systems. To tackle this challenge, a promising new approach involves combining hybrid quantum-classical neural networks with geo-information technologies. Methods. The study utilized quantum-enhanced hybrid neural networks in combination with GIS methods to identify named entities such as personal accounts and balance sheet objects of Kyivvodokanal by their addresses and geocoding. This information was then published on a geoportal using the ArcGIS Enterprise platform in real-time, which holds great promise for effective water management. The performance of the developed model was evaluated by accuracy indicators, recall parameters, and weighted harmonic average of accuracy and recall. Results. The obtained results indicate that the developed hybrid quantum-classical model of artificial intelligence can be successfully applied to transform large volumes of unstructured textual information into spatial information. The model was integrated into GIS using ArcGIS Enterprise. By combining the obtained point classes of spatial objects with already existing data, methods of spatial connections, an interactive map with an update interval of every five minutes was developed. Conclusions. Taking advantage of quantum computing and combining it with classical hardware and classical AI models, it became possible to achieve similar and even better performance in various tasks compared to state-of-the-art methods. Quantum natural language processing is a promising new field that has the potential to revolutionize the way one analyzes and understands human language.

������� ����������-����ֲ��������� ������� �������ֲ���-���˲����ί ������� ��� ϲ������� �� ��������� в���� ���� ��ò�������� ������������ ��ֲ������ί ����IJ� �������

The article establishes that the implementation of an information and analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine is possible only if a single information space of the Ministry of Internal Affairs of Ukraine (as a metasystem) is formed. Such an information-analytical system will ensure the full-scale integrated use of distributed information resources and corporate knowledge systems that are or may be relevant to the processes of logistics support of the National Guard of Ukraine. It is shown that the functional services of the information-analytical system for support and decision-making on the logistics support of the National Guard of Ukraine should include two groups of transdisciplinary procedures, on the basis of which the component services of the information-analytical system are formed: these are the services of linguistic and semantic processing of information resources and transdisciplinary analysis and decision support. It has been established that the ontological-semantic group of services of the information-analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine includes transdisciplinary services that provide processing and appropriate transformations of unstructured textual information. The composition, content, and purpose of the ontological-semantic group of services of the information-analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine are determined. It is established that the ontological-analytical group includes the services of the information-analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine for the analytical activities of experts. The composition, content and purpose of the ontological-analytical group of services of the information-analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine are determined. Taking into account the essence, methods of implementation and specifics of the polyhedron processes considered in the article, the main recommendations for choosing the component configuration of the polyhedron processes of the information-analytical system for supporting and making decisions on the logistics support of the National Guard of Ukraine are developed.

Analysis of approaches to the semantic processing of unstructured textual information

Scientific and methodological approaches to semantic processing of unstructured text information in computer systems based on the methodology of heuristic and neuromathematical data processing are considered. The results of a conceptual analysis of heuristic and neuromathematical methods for semantic processing of unstructured text information are presented. The basic provisions for the implementation of semantic processing of unstructured text information based on the integrated application of heuristic and neuromathematical approaches are outlined.

A System for Converting and Recovering Texts Managed as Structured Information

This paper introduces a system that incorporates several strategies based on scientific models of how the brain records and recovers memories. Methodologically, an incremental prototyping approach has been applied to develop a satisfactory architecture that can be adapted to any language. A special case is studied and tested regarding the Spanish language. The applications of this proposal are vast because, in general, information such as text way, reports, emails, and web content, among others, is considered unstructured and, hence, the repositories based on SQL databases usually do not handle this kind of data correctly and efficiently. The conversion of unstructured textual information to structured one can be useful in contexts such as Natural Language Generation, Data Mining, and dynamic generation of theories, among others.

Integration of Sentiment Analysis of Social Media in the Strategic Planning Process to Generate the Balanced Scorecard

Strategic planning (SP) requires attention and constant updating and is a crucial process for guaranteeing the efficient performance of companies. This article proposes a novel approach applied in a case study whereby a balanced scorecard (BSC) was generated that integrated sentiment analysis (SA) of social media (SM) and took advantage of the valuable knowledge of these sources. In this study, opinions were consolidated in the main dataset to incorporate sentiments regarding the strategic part of a restaurant in a tourist city. The proposed methodology began with the selection of the company. Information was then acquired to apply pre-processing, processing, evaluation, and validation that is capitalized in a BSC to support strategic decision-making. Python support was used in the model and comprised lexicon and machine learning approaches for the SA. The significant knowledge in the comments was automatically oriented toward the key performance indicators (KPIs) and perspectives of a BSC that were previously determined by a group of opinion leaders of the company. The methods, techniques, and algorithms of SA and SP showed that unstructured textual information can be processed and capitalized efficiently for optimal management and decision-making. The results revealed an improvement (reduced effort and time) to produce a more robust and comprehensive BSC with the support and validation of experts. Moreover, new resources and approaches were developed to implement more efficient SP. The model was based on the efficient coupling of both fields of study.

Global meta-analysis of evolution patterns for lake topics over centurial scale: A natural language understanding-based deep clustering approach with 130,000 studies

As complicated microcosms, lakes have been attracting exponentially increasing attention, resulting in plentiful interdisciplinary academic publications of more than 105. It is thereby challengeable to explore the massive unstructured text information of publications to understand lake topics from the global- and centurial-scale perspectives. However, conventional bibliometrics suffer from the limitations of non-understanding the literature. A novel approach, Natural Language Understanding-Based Deep Clustering (NLU-DC) for large text clustering, was proposed in this study for global meta-analysis of evolution patterns for lake topics. The validated NLU-DC elevated the available keywords from 24% to 70%, correcting the statistical bias in the traditional evidence synthesis. Its high performance derives from the integration of a deep learning model, cosine distance, DBSCAN clustering and changing hyperparameters. This approach is of great accuracy and efficiency for large text datasets. This study thereby identified the centurial-scale topics related to lakes using large literature datasets covering >130,000 studies. The results showed that the topics became more and more abundant but were concentrated stably towards central ones. Six evolution patterns, consisting of fluctuating, emerging in 1970, emerging after 1970, trending-up, stable and trending-down patterns, were identified with generalized linear model (GLM). We found that, in recent twenty years, few emerging topics attract significant academic attention; and the dependency between topics is catching more attention than before. To prolong the lake studies, it is essential to strengthen the integrated studies on multi-pattern topics, in particular, over emerging and trending-down topics. Our study verified that the NLU-DC, consisting of state-of-the-art deep learning models in natural language processing and efficient clustering algorithm in machine learning, is a powerful method for global meta-analysis of water-related research fields and have huge potential to be applied in all fields of academical studies.

RECA: Relation Extraction Based on Cross-Attention Neural Network

Extracting entities and relations, as a crucial part of many tasks in natural language processing, transforms the unstructured text information into structured information and provides corresponding data support for knowledge graph (KG) and knowledge vault (KV) construction. Nevertheless, the mainstream relation-extraction methods, the pipeline method and the joint method, ignore the dependency between the subject entity and the object entity. This work introduces a pre-trained BERT model and a dilated gated convolutional neural network (DGCNN) as an encoder to distinguish the long-range semantics representation from the input sequence. In addition, we propose a cross-attention neural network as a decoder to learn the importance of each subject word for each word of the input sequence. Experiments were undertaken with two extensive datasets, the New York Times Corpus (NYT) and WebNLG Corpus, and showed that our model performs significantly better than the CasRel model, outperforming the baseline by 1.9% and 0.7% absolute gain in terms of F1-score.

CitEnergy : A BERT based model to analyse Citizens’ Energy-Tweets

Micro-blogging social site Twitter has emerged as a rich source of unstructured text information which could be processed and analysed to extract people’s opinions about several topics and events including natural hazards, energy, sports, transportation, elections etc. The present research study adds a novel perspective in this dimension by extracting citizens’ opinions on several electricity-related issues. Recent research studies in this domain have employed Bag-of-Words (BoW) model for the numerical representation of tweets. However, the BoW model suffers from several issues such as incapable of handling the semantic relationship between words, sparse and high dimensional representation. The present research work overcomes all aforementioned shortcomings by integrating popular word-embeddings with deep learning models for the classification tasks. The study harnesses social media data for two classification tasks: Sentiment classification task and Complaints classification task. Firstly, a series of preprocessing steps are applied on tweets extracted from Twitter streaming API. Subsequently, different word-embeddings models are employed to generate a numerical representation of tweets while capturing the semantic relationship among words. Several deep learning-based sentiment classification models are then deployed on top of generated word-embeddings for identifying/classifying citizens’ sentiments from the tweets. Lastly, the tweets associated with negative sentiment class (identified by the sentiment classification model) are further processed and analysed for building the complaints classification model. The complaints classification model prioritize and assign negative sentiment tweets into one of the two target classes depending on the target issue raised in the tweets (Community level or recurring complaint and Individual level complaint). In addition to this, the current study also proposes Bi-directional Encoded Representations for Transformers (BERT) based Sentiment classification and Complaints classification models for achieving the improved classification accuracy. Experimental evaluation of the proposed BERT based models is done by comparing prediction results with several benchmark deep learning models.