Natural Language Processing Approaches Research Articles

Analyzing social media reactions to the FTX crisis: Unraveling the spillover effect on crypto markets

AbstractIn this study, we analyzed social media conversations during the unfolding of the FTX crisis, the biggest cryptocurrency scandal in United States history. Drawing on the accessibility‐diagnosticity framework, we examined the negative spillover effect of the crisis using a natural language processing approach. We specifically assessed whether there was a negative spillover from FTX to other crypto entities with different levels of diagnostic attribute similarity. We collected a large corpus of Twitter conversations related to the FTX collapse in 2022 and used the association rule analysis to determine the association between FTX and other crypto entities. Our analysis revealed that the number of tweets mentioning FTX and other crypto entities changed in line with a series of real‐world events during the FTX crisis. The negative spillover of the FTX crisis occurred primarily during the first 10 days as the FTX scandal unfolded. The results indicated that the FTX crisis spilled over to highly accessible and diagnostic crypto entities, such as Binance, Bitcoin, and the cryptocurrency industry in general. On the other hand, less accessible and less diagnostic crypto entities/currencies like Ethereum and Coinbase did not experience negative spillover from the scandal.

Validation of a Natural Language Processing Algorithm using National Reporting Data to Improve Identification of Anesthesia-related ADVerse evENTs: The “ADVENTURE” Study

BackgroundReporting and analysis of adverse events (AE) is associated with improved healthcare learning, quality outcomes, and patient safety. Manual text analysis is time-consuming, costly, and prone to human errors. We aimed to demonstrate the feasibility of machine learning and natural language processing (NLP) approaches for early predictions of adverse events and provide input to direct quality improvement and patient safety initiatives. MethodsWe used machine learning to analyze 9,559 continuously reported AE by clinicians and healthcare systems to the French National Health accreditor (HAS) between January 1, 2009, and December 31, 2020, for a total of 135,000 unique de-identified AE reports. We validated the labeling and determined the associations between different root causes and patient consequences. The model was validated by independent expert anesthesiologists. ResultsThe machine learning and Artificial Intelligence (AI) model trained on 9,559 AE datasets accurately categorized 8800 (88%) of reported AE. The three most frequent AE types were “difficult orotracheal intubation” (16.9% of AE reports), “medication error” (10.5%), and “post-induction hypotension” (6.9%). The accuracy of the AI model reached 70.9% sensitivity, 96.6% specificity for “difficult intubation”, 43.2% sensitivity, and 98.9% specificity for “medication error.” ConclusionsThis unsupervised method provides an accurate, automated, AI-supported search algorithm that ranks and helps to understand complex patterns of risky patient situations and has greater speed, precision, and clarity when compared to manual human data extraction. Machine learning (ML) and natural language processing models can effectively be used to process natural language AE reports and augment expert clinician input. This model can support clinical applications and methodological standards of implementations and be used to better inform and enhance decision-making for improved risk management and patient safety. Trial RegistrationThe study was approved by the ethics committee of the French Society of Anesthesiology (IRB 00010254-2020-20) and the CNIL (CNIL: 118 58 95) and the study was registered with ClinicalTrials.gov (NCT: NCT05185479).

Building large-scale registries from unstructured clinical notes using a low-resource natural language processing pipeline

Building clinical registries is an important step in clinical research and improvement of patient care quality. Natural Language Processing (NLP) methods have shown promising results in extracting valuable information from unstructured clinical notes. However, the structure and nature of clinical notes are very different from regular text that state-of-the-art NLP models are trained and tested on, and they have their own set of challenges. In this study, we propose Sentence Extractor with Keywords (SE-K), an efficient and interpretable classification approach for extracting information from clinical notes and show that it outperforms more computationally expensive methods in text classification. Following the Institutional Review Board (IRB) approval, we used SE-K and two embedding based NLP approaches (Sentence Extractor with Embeddings (SE-E) and Bidirectional Encoder Representations from Transformers (BERT)) to develop comprehensive registry of anterior cruciate ligament surgeries from 20 years of unstructured clinical data at a multi-site tertiary-care regional children's hospital. The low-resource approach (SE-K) had better performance (average AUROC of 0.94 ± 0.04) than the embedding-based approaches (SE-E: 0.93 ± 0.04 and BERT: 0.87 ± 0.09) for out of sample validation, in addition to minimum performance drop between test and out-of-sample validation. Moreover, the SE-K approach was at least six times faster (on CPU) than SE-E (on CPU) and BERT (on GPU) and provides interpretability. Our proposed approach, SE-K, can be effectively used to extract relevant variables from clinic notes to build large-scale registries, with consistently better performance compared to the more resource-intensive approaches (e.g., BERT). Such approaches can facilitate information extraction from unstructured notes for registry building, quality improvement and adverse event monitoring.

Does the Gender of Doctors Change a Patient’s Perception?

ABSTRACT In the past, researchers have used various data sources like social media, admission applications, or letters of recommendation to identify gender-based differences in linguistic data. One such avenue in healthcare is the online physician reviews website. Such websites, for example, RateMyDoctors.com, ZocDoc, or even Yelp.com have become a go-to place for patients when choosing their physicians. In the current research, we used two different natural language processing (NLP) approaches: semi-supervised and unsupervised topic modeling to analyze the text of the reviews to identify gender-based linguistic differences from patients’ perspectives. We found that female physicians receive more reviews on their personable skills and warmth, aligning with the Stereotype Content Model. We also found other popular topics discussing bedside manners and overall patient experiences, where the reviews suggested that patients were happier with their experience with female physicians and perceived them to have more positive traits than their male counterparts. Although our study did not reflect significant linguistic differences; it highlights the importance for patients and doctors to be more aware of potential gender stereotypes and perceptions.

Assessment of Gender Differences in Letters of Recommendation for Physical Therapy Residency Applications.

Letters of recommendation (LOR) are an integral component of physical therapy residency applications. Identifying the influence of applicant and writer gender in LOR will help identify whether potential implicit gender bias exists in physical therapy residency application processes. Several medical and surgical residency education programs have reported positive, neutral, or negative LOR female gender bias among applicants and writers. Little research exists on gender differences in LOR to physical therapy education programs or physical therapy residency programs. Seven hundred sixty-eight LOR were analyzed from 256 applications to 3 physical therapy residency programs (neurologic, orthopaedic, sports) at one institution from 2014 to 2020. Thematic categories were developed to identify themes in a sample of LOR. Associations between writer and applicant gender were analyzed using summary statistics, word counts, thematic and psycholinguistic extraction, and rule-based and deep learning Natural Language Processing . No significant difference in LOR word counts were found based on writer or applicant gender. Increased word counts were seen in sports residency LOR compared with the orthopaedic residency. Thematic analysis showed LOR gender differences with male applicants receiving more positive generalized recommendations and female applicants receiving more comments regarding interpersonal relationship skills. No thematic or psycholinguistic gender differences were seen by LOR writer. Male applicants were 1.9 times more likely to select all male LOR writers, whereas female applicants were 2.1 times more likely to choose all female LOR writers. Gender differences in LORs for physical therapy residencies were found using a comprehensive Natural Language Processing approach that identified both a positive recommendation male applicant gender bias and a positive interpersonal relationship skill female applicant gender bias. Applicants were not harmed nor helped by selecting LOR writers of the opposite gender. Admissions committees and LOR writers should be mindful of potential implicit gender biases in LOR submitted to physical therapy residency programs.

Decoding violence against women: analysing harassment in middle eastern literature with machine learning and sentiment analysis

The rising prevalence of harassment in Middle Eastern countries is mirrored in literary works from the region. However, extracting data from these texts to understand the typology and frequency of the cases poses a significant challenge due to human cognitive limitations and potential biases. Thus, this study aims to use natural language processing (NLP) approaches to propose a machine learning framework for text mining of sexual harassment content in literary texts. The data source for this study consists of twelve Middle Eastern novels. The proposed framework involves the classification of physical and non-physical types of sexual harassment using a machine-learning model. Lexicon-based sentiment and emotion detection are applied to sentences containing instances of sexual harassment for data labelling and analysis. Finally, a long short-term memory-gated recurrent unit (LSTM-GRU) deep learning model is built to classify the sentiment characteristics that induce sexual harassment. The proposed model achieved an accuracy of 75.8% while outperforming five other algorithms. Additionally, a sentiment classification with three labels—negative, positive, and neutral—was developed using an LSTM-GRU RNN deep learning model. The accuracy of this model was 84.5%. Most statements, even those involving physical sexual harassment, which had greater levels of sexual harassment, had negative sentiments, according to lexicon-based sentiment analysis. This study contributes to the field of text mining by providing a novel approach to identifying instances of sexual harassment in literature in English from the Middle East. The use of machine learning models and sentiment analysis techniques allows for more accurate identification and classification of different types of sexual harassment. Furthermore, this study sheds light on the prevalence of sexual harassment in Middle Eastern countries and highlights the need for further research and action to address this issue.

Enhancing Financial Sentiment Analysis: A Deep Dive into Natural Language Processing for Market Prediction Industries

The purpose of this study is to investigate the enhancement of Financial Sentiment Analysis by conducting an in-depth investigation of Natural Language Processing (NLP) approaches for the purpose of improving market prediction. The purpose of this research is to investigate the potential of natural language processing (NLP) to improve the accuracy and efficiency of sentiment analysis. This is in response to the complex structure of financial markets and the crucial role that sentiment plays. The examination of the relevant literature highlights the limits of traditional methods and the urgent need for creative solutions in the field of financial sentiment research. The approach that we use entails the careful collecting of data from social media and financial news, with a particular emphasis on the utilization of strong pre-processing tools. The research assesses the performance parameters of accuracy, precision, recall, and correlation with market trends by using natural language processing (NLP) technologies such as algorithms for sentiment analysis, Named Entity Recognition, and deep learning models. The findings include a comparative examination of conventional methods and those based on natural language processing (NLP), therefore revealing insights into the significant influence that sentiment has on market patterns. The results not only provide a contribution to the theoretical knowledge of sentiment research, but they also offer real consequences for financial analysts who are looking to make market forecasts that are more accurate and timelier. The research suggests ways for refinement, with an emphasis on enhanced pre-processing and Explainable AI integration. These tactics are being proposed to address issues in data quality and bias. When looking to the future, the study provides an overview of potential future paths, which include the investigation of external influences and the development of deep learning models for accurate market forecasting respectively. To summaries, the findings of this research establish natural language processing (NLP) as a revolutionary force in the process of redefining financial sentiment analysis. Furthermore, it offers a path for future developments in the ever-changing world of market prediction.

Unveiling Gender Biases in Recruitment: A Natural Language Processing Approach

This paper investigates the potential of AI to identify gender biases in recruitment for senior management positions in businesses, dealing with many documents. It aims to unravel the impact of gender biases in job advertisements as a possible reason behind the underrepresentation of women in the corporate world. An innovative experiment that extracts and analyses 2.198 job offers published in February and September 2021 in the Financial Times newspaper is presented. Natural language techniques are used. These methods identify the most frequent terms and their appearance rate in the advertisements showing the gender biases they generate. By enabling the analysis of many documents, the method allows the accurate identification of gender biases. This use is unique in management studies. The results show a strong co-occurrence of terms associated with male roles in the studied sectors. The concept of agentic-communal role differentiation, rooted in the Identity, homosocial, and TM-TM theories, supports the findings. This knowledge will contribute to using NLP to discover gender biases in recruitment for high decision-making positions proposing action to improve the present situation. Should more women ascend to decision-making positions, selection processes should be improved to reflect a more neutral language. At the same time, cultural changes should be promoted in the corporate world toward more inclusive workplaces.

Social Media and Real Estate: Do Twitter Users Predict REIT Performance?

This study investigates the impact of social media sentiment on indirect real estate market returns by utilizing a comprehensive natural language processing approach to identify relevant Twitter posts and extract sentiment from them. To handle the complex linguistic features inherent in social media messages, three different sentiment classifiers are compared. The findings suggest a significant relationship between monthly sentiment and REIT returns, which occurs in two phases: a short-term speculative reaction and a greater longer-term reaction related to actual changes in the real estate market. The study also highlights that while the conventional dictionary approach can identify this relationship, more sophisticated classifiers can achieve higher accuracy. Overall, the results demonstrate the valuable insights that can be gained from analyzing social media data and its potential impact on the real estate market.

An English Translation Error Correction System Based on Multi-Feature Fusion

Multi-feature fusion is an advanced computational technique that combines multiple sources of information or characteristics to improve the performance of a system or model. In various fields such as computer vision, natural language processing, and machine learning, this approach integrates diverse features like visual cues, textual data, and contextual information to enhance accuracy, robustness, and efficiency. Multi-feature fusion-based English translation represents a cutting-edge approach in natural language processing, where diverse linguistic features are amalgamated to refine the accuracy and quality of translation outcomes. By integrating various elements such as syntax, semantics, context, and linguistic patterns, this method significantly enhances the translation process, ensuring more nuanced and contextually precise results. This paper introduces an innovative English Translation Error Correction System (ETECS) founded on Multi-Feature Fusion, supplemented by a bi-gram Statistical Classification Neural Network (n-gramCNN). In response to the growing demand for accurate and efficient translation error detection and correction, this research endeavors to enhance the quality of English translations through a comprehensive and advanced methodology. Using a dataset of 1,000 English text samples with corresponding translations, ETECS demonstrates an impressive error detection accuracy of 95%. Upon identifying errors, the system achieves a correction accuracy of 92%, effectively rectifying grammatical, syntactical, and semantic inconsistencies. Moreover, ETECS outperforms baseline error correction systems by a significant margin, showcasing its superiority in handling complex translation errors. The bi-gram Statistical Classification Neural Network (n-gramCNN) contributes to this success by achieving an F1 score of 0.94, highlighting its effectiveness in classifying and correcting translation errors.

Regular paper Comparative Analysis of Deep Learning Models for Sentiment Analysis on IMDB Reviews

Within the domain of natural language processing, sentiment analysis assumes a fundamental position, facilitating the comprehension of societal perspectives and opinions, thereby playing a pivotal role in understanding public sentiment. In this study, an examination and comparison of deep learning architectures was conducted on IMDB movie reviews. We evaluated the performance of Basic Recurrent Neural Networks, Long Short-Term Memory (LSTM), Gated Recurrent Neural Unit (GRU), Bidirectional LSTM, Bidirectional GRU and 1D Convolutional Neural Networks (Conv1D) based on their training, validation, and testing accuracies. Our results indicate that while LSTM achieved the highest accuracy of 99.91% on the training data, GRU demonstrated superior performance (88.28%) on the validation dataset. Interestingly, Bidirectional GRU emerged as the top-performing model (87.54%) on the testing data, showcasing its robustness in generalizing to unseen instances. These findings highlight the importance of evaluating model performance across multiple datasets to assess their real-world effectiveness. Furthermore, our comparative analysis provides valuable understanding of the advantages and limitations of each model, offering practical guidance for selecting the optimal framework for sentiment analysis endeavors. Overall, this research contributes to the progress of such methodologies and deep learning approaches in natural language processing.

227 Spine-tuned Natural Language Models and Bespoke Regular Expression Classifiers for Automated Spinal Surgery Registry Development

INTRODUCTION: Surgical research demands the development of clinical registries, often through time-intensive manual chart review. Natural language processing (NLP) may accelerate registry development, and an ideal automatic registry (autoregistry) algorithm would be highly accurate while requiring minimal manual data annotation. NLP approaches including bespoke Regular Expression (RegEx) classifiers and Large Language Models (LLM) possess distinct strengths and weaknesses and have not been compared in the setting of autoregistry development. METHODS: We used an institutional data lake to retrieve 31,502 neurosurgical operative notes. A standardized set of spinal procedures was chosen for inclusion in the autoregistry. 200 manually annotated notes were used for training and testing purposes. RegEx classifiers were engineered to retrieve procedural info from unprocessed notes. A family of 110-million parameter BERT models, including LLM pre-trained on clinical text, was fine-tuned for the same tasks. We also tested a open-source 7-billion parameter LLM chatbot, Vicuna, without fine-tuning. RESULTS: The RegEx classifiers were able to identify spinal procedures and associated vertebral levels in nearly 99% of operative notes. Fine-tuned LLM identified common procedures (e.g. spinal fusion and laminectomy) with greater than 95% accuracy but performed poorly for rarer procedures (e.g. XLIF, corpectomy) and vertebral body identification. Qualitative evaluation of the Vicuna chatbot showed potential for the same tasks, following iteratively refined prompting. CONCLUSIONS: The goal of autoregistry development is to minimize time- and labor-intensive manual chart review. We found that fine-tuned LLM could not match the accuracy and efficiency of the RegEx classifier. However, LLM may be well-suited to expand existing clinical databases that provide a robust training set. Further work combining NLP approaches will attempt to develop a pipeline for autoregistry development from natural language (plain English) queries.

Leveraging machine learning and NLP for enhanced cohorting and RxNorm mapping in Electronic Health Records (EHRs)

This work addresses the combination of machine learning (ML) and natural language processing (NLP) approaches to optimize the process of courting and RxNorm mapping inside Electronic Health Records (EHRs). Cohorting patients based on comparable traits or diseases is vital for clinical research, but it generally depends on time-consuming manual techniques and is prone to mistakes. Similarly, mapping pharmaceutical names to standardized codes such as RxNorm promotes interoperability and data analysis but may be challenging owing to variances in how drugs are reported. Leveraging ML and NLP may automate and optimize these procedures, leading to more efficient cohort identification and precise medication mapping. We offer a thorough technique for integrating ML and NLP algorithms in EHR systems, including data preparation, feature engineering, model training, and assessment. Through testing and analysis, we show the usefulness of our technique in enhancing cohorting accuracy and RxNorm mapping precision. The findings underline the promise of ML and NLP in revolutionizing EHR data management, leading to improved patient care and simplified research procedures.

Maternal mental health monitoring in an online community: a natural language processing approach

ABSTRACT Digital maternity support communities are increasingly popular. The communities are often based on discussion forums called ‘birth clubs’, to which users are assigned according to their estimated due months. Distinguishing between support-seeking and non-support-seeking posts submitted to these ‘birth clubs’ is a crucial first step for monitoring maternal mental health. This study utilised natural language processing (NLP) techniques on 52,558 posts collected from one of the largest online maternity communities in China, employing machine learning algorithms trained for post classification with a randomly selected and manually labelled subset of 3000 posts. The results validated the properties of information similarity and time sensitivity within the post data, and demonstrated the feasibility of employing simple algorithms and small training sets for effective maternal mental health monitoring.

Assessing the utility of natural language processing for detecting postoperative complications from free medical text.

Postoperative complication rates are often assessed through administrative data, although this method has proven to be imprecise. Recently, new developments in natural language processing have shown promise in detecting specific phenotypes from free medical text. Using the clinical challenge of extracting four specific and frequently undercoded postoperative complications (pneumonia, urinary tract infection, sepsis, and septic shock), it was hypothesized that natural language processing would capture postoperative complications on a par with human-level curation from electronic health record free medical text. Electronic health record data were extracted for surgical cases (across 11 surgical sub-specialties) from 18 hospitals in the Capital and Zealand regions of Denmark that were performed between May 2016 and November 2021. The data set was split into training/validation/test sets (30.0%/48.0%/22.0%). Model performance was compared with administrative data and manual extraction of the test data set. Data were obtained for 17 486 surgical cases. Natural language processing achieved a receiver operating characteristic area under the curve of 0.989 for urinary tract infection, 0.993 for pneumonia, 0.992 for sepsis, and 0.998 for septic shock, whereas administrative data achieved a receiver operating characteristic area under the curve of 0.595 for urinary tract infection, 0.624 for pneumonia, 0.571 for sepsis, and 0.625 for septic shock. The natural language processing approach was able to capture complications with acceptable performance, which was superior to administrative data. In addition, the model performance approached that of manual curation and thereby offers a potential pathway for complete real-time coverage of postoperative complications across surgical procedures based on natural language processing assessment of electronic health record free medical text.

Advancing Job Search: A Comprehensive Resume-Based Job Recommendation System Using NLP and Deep Learning Techniques

Abstract: Students find it difficult to find jobs and postings related to their skills and which has a good pay. Students aimlessly scan all internet to find a job which best matches their skill. Hence this paper presents a model which would give them personalized recommendations based on their skills mentioned in the resume. In today’s world hundreds of people apply for a single job posting. Companies receive thousands of applications at any given time. To save time companies, recruiters have few seconds to go through the applicant’s resume. Moreover, today all companies use a software based application which scans all resumes and get the best candidates. Most students especially freshers fail at this critical juncture, they fail to make a good resume which would help their resumes getting selected. Our novel model helps solve this problem. In order to extract sectionspecific text content from resumes, this work presents a layout-aware resume parsing system based on natural language processing (NLP) and rule-based approaches. This output can be fed into a resume content review system to obtain resume feedback, or it can be utilized as the input for a resume content score model. By utilizing sophisticated natural language processing (NLP) models, the system guarantees precise recognition and classification of essential resume elements, including personal information, qualifications, experience, and accomplishments. The recommendation engine makes use of machine learning techniques to pinpoint areas that need work, such improving keyword relevancy, recommending extra abilities, or offering formatting and wording advice. The goal of this iterative process is to provide job seekers with dynamic tools that will boost their visibility to recruiters and applicant tracking systems. Our unique model also recommends skills to add in their resume. This models also recommends some add-on courses to improve those skills. The app also uses cutting-edge technology like Natural Language Processing (NLP) and Optical Character Recognition (OCR) to process resumes and job listings and find the greatest fit for both job seekers.

The Optimization of a Natural Language Processing Approach for the Automatic Detection of Alzheimer's Disease Using GPT Embeddings.

The development of noninvasive and cost-effective methods of detecting Alzheimer's disease (AD) is essential for its early prevention and mitigation. We optimize the detection of AD using natural language processing (NLP) of spontaneous speech through the use of audio enhancement techniques and novel transcription methodologies. Specifically, we utilized Boll Spectral Subtraction to improve audio fidelity and created transcriptions using state-of-the-art AI services-locally-based Wav2Vec and Whisper, alongside cloud-based IBM Cloud and Rev AI-evaluating their performance against traditional manual transcription methods. Support Vector Machine (SVM) classifiers were then trained and tested using GPT-based embeddings of transcriptions. Our findings revealed that AI-based transcriptions largely outperformed traditional manual ones, with Wav2Vec (enhanced audio) achieving the best accuracy and F-1 score (0.99 for both metrics) for locally-based systems and Rev AI (standard audio) performing the best for cloud-based systems (0.96 for both metrics). Furthermore, this study revealed the detrimental effects of interviewer speech on model performance in addition to the minimal effect of audio enhancement. Based on our findings, current AI transcription and NLP technologies are highly effective at accurately detecting AD with available data but struggle to classify probable AD and mild cognitive impairment (MCI), a prodromal stage of AD, due to a lack of training data, laying the groundwork for the future implementation of an automatic AD detection system.

Automation in Library Science in AI Era

The Auromation of library science has become a revolutionary force in the adminstration and accessibility of information resources in the age of artifical intelliegence. The goal of this project is to investigate and apply AI -driven solution to improve resource management, user services, cataloguing and other areas of library operation. in order to streamline the organisation and retrieval of information, the automation process uses machine learning algorithms to tag and categorise library material. Intelligent search and recommendation systems are developed using Natural Language Processing(NLP ) approaches, which facilitities the effortless discovery of pertient information by users. Additionally, by integrating AI into library systems, libraries may better forecast demand using predictive analytics and adapt their resources accourding to user perfence and new trends. Robotic process automation also simplifies repetitive administrative work, freeing up human resources to concentrate on more difficult and valuable duties. The project uses strong encryption techiques and access control to address issue with data security and privacy, the creation and application of AI technologies in library science also heavily relies on ethical consideration, such as bias prevention in algorithms. Libraries may transform their services and remain ata the forefront of information management technology by adopting automation in the AI era and offering users a more efficient and customised experience. This abstract highlights the potential for major breakthroughs in the field of library science through the integration of artifical intellience , outlining the project's main goal and technique.

AssistMED project: Transforming cardiology cohort characterisation from electronic health records through natural language processing – Algorithm design, preliminary results, and field prospects

IntroductionElectronic health records (EHR) are of great value for clinical research. However, EHR consists primarily of unstructured text which must be analysed by a human and coded into a database before data analysis- a time-consuming and costly process limiting research efficiency. Natural language processing (NLP) can facilitate data retrieval from unstructured text. During AssistMED project, we developed a practical, NLP tool that automatically provides comprehensive clinical characteristics of patients from EHR, that is tailored to clinical researchers needs. Material and methodsAssistMED retrieves patient characteristics regarding clinical conditions, medications with dosage, and echocardiographic parameters with clinically oriented data structure and provides researcher-friendly database output. We validate the algorithm performance against manual data retrieval and provide critical quantitative and qualitative analysis. ResultsAssistMED analysed the presence of 56 clinical conditions, medications from 16 drug groups with dosage and 15 numeric echocardiographic parameters in a sample of 400 patients hospitalized in the cardiology unit. No statistically significant differences between algorithm and human retrieval were noted. Qualitative analysis revealed that disagreements with manual annotation were primarily accounted to random algorithm errors, erroneous human annotation and lack of advanced context awareness of our tool. ConclusionsCurrent NLP approaches are feasible to acquire accurate and detailed patient characteristics tailored to clinical researchers' needs from EHR. We present an in-depth description of an algorithm development and validation process, discuss obstacles and pinpoint potential solutions, including opportunities arising with recent advancements in the field of NLP, such as large language models.

Measuring and Mitigating Gender Bias in Legal Contextualized Language Models

Transformer-based contextualized language models constitute the state-of-the-art in several natural language processing (NLP) tasks and applications. Despite their utility, contextualized models can contain human-like social biases, as their training corpora generally consist of human-generated text. Evaluating and removing social biases in NLP models has been a major research endeavor. In parallel, NLP approaches in the legal domain, namely, legal NLP or computational law, have also been increasing. Eliminating unwanted bias in legal NLP is crucial, since the law has the utmost importance and effect on people. In this work, we focus on the gender bias encoded in BERT-based models. We propose a new template-based bias measurement method with a new bias evaluation corpus using crime words from the FBI database. This method quantifies the gender bias present in BERT-based models for legal applications. Furthermore, we propose a new fine-tuning-based debiasing method using the European Court of Human Rights (ECtHR) corpus to debias legal pre-trained models. We test the debiased models’ language understanding performance on the LexGLUE benchmark to confirm that the underlying semantic vector space is not perturbed during the debiasing process. Finally, we propose a bias penalty for the performance scores to emphasize the effect of gender bias on model performance.