Search

We document and explain the sharp performance deterioration of smart beta indexes after the corresponding exchange-traded funds (ETFs) are launched for investment. While smart beta is purported to deliver excess returns through factor exposures, the market-adjusted return of smart beta indexes drops from about 3% "on paper" before ETF listings to about −0.50% to −1% after ETF listings. This performance decline cannot be explained by variation in factor premia, strategic timing, or diminishing returns to scale. Instead, we find strong evidence of data mining in the construction of smart beta indexes, which helps ETFs attract flows, as investors respond positively to backtests. [ABSTRACT FROM AUTHOR]

The increasing development of the competitive market has forced organisations to make great efforts in the processes of supply, production, and distribution to meet customer demand in the shortest time and at the lowest cost. A cross-docking (CD) system is one of the successful and practical strategies in this field considered by researchers in various fields. Also, business process management plays a key role in continuous improvement and increased productivity. In today's digital age, due to the ability to record all activities, process mining is an important method to identify the current situation and improve productivity. In this research, a newly established CD belonging to a chain store is studied to improve the current situation, in which different goods enter and then exit after different processes. The purpose of this study is to obtain the optimal number of doors and loaders as sources. First, helping an RFID system, all activities are recorded, and the current situation of the processes is monitored, and then, the real process model is identified using heuristic and inductive miner algorithms. After adapting to the event log by using the simulation process in Arena software, different scenarios are examined, and the best possible case is presented. [ABSTRACT FROM AUTHOR]

In the United States, hundreds of thousands of undocumented orphan wells have been abandoned, leaving the burden of managing environmental hazards to governmental agencies or the public. These wells, a result of over a century of fossil fuel extraction without adequate regulation, lack basic information like location and depth, emit greenhouse gases, and leak toxic substances into groundwater. For most of these wells, basic information such as well location and depth is unknown or unverified. Addressing this issue necessitates innovative and interdisciplinary approaches for locating, characterizing, and mitigating their environmental impacts. Our survey of the United States revealed the need for tools to identify well locations and assess conditions, prompting the development of technologies including machine learning to automatically extract information from old records (95%+ accuracy), remote sensing technologies like aero-magnetometers to find buried wells, and cost-effective methods for estimating methane emissions. Notably, fixed-wing drones equipped with magnetometers have emerged as cost-effective and efficient for discovering unknown wells, offering advantages over helicopters and quadcopters. Efforts also involved leveraging local knowledge through outreach to state and tribal governments as well as citizen science initiatives. These initiatives aim to significantly contribute to environmental sustainability by reducing greenhouse gases and improving air and water quality.

ObjectiveAlthough supervised machine learning is popular for information extraction from clinical notes, creating large annotated datasets requires extensive domain expertise and is time-consuming. Meanwhile, large language models (LLMs) have demonstrated promising transfer learning capability. In this study, we explored whether recent LLMs could reduce the need for large-scale data annotations.Materials and methodsWe curated a dataset of 769 breast cancer pathology reports, manually labeled with 12 categories, to compare zero-shot classification capability of the following LLMs: GPT-4, GPT-3.5, Starling, and ClinicalCamel, with task-specific supervised classification performance of 3 models: random forests, long short-term memory networks with attention (LSTM-Att), and the UCSF-BERT model.ResultsAcross all 12 tasks, the GPT-4 model performed either significantly better than or as well as the best supervised model, LSTM-Att (average macro F1-score of 0.86 vs 0.75), with advantage on tasks with high label imbalance. Other LLMs demonstrated poor performance. Frequent GPT-4 error categories included incorrect inferences from multiple samples and from history, and complex task design, and several LSTM-Att errors were related to poor generalization to the test set.DiscussionOn tasks where large annotated datasets cannot be easily collected, LLMs can reduce the burden of data labeling. However, if the use of LLMs is prohibitive, the use of simpler models with large annotated datasets can provide comparable results.ConclusionsGPT-4 demonstrated the potential to speed up the execution of clinical NLP studies by reducing the need for large annotated datasets. This may increase the utilization of NLP-based variables and outcomes in clinical studies.

Advanced omics technologies and facilities generate a wealth of valuable data daily; however, the data often lack the essential metadata required for researchers to find, curate, and search them effectively. The lack of metadata poses a significant challenge in the utilization of these data sets. Machine learning (ML)-based metadata extraction techniques have emerged as a potentially viable approach to automatically annotating scientific data sets with the metadata necessary for enabling effective search. Text labeling, usually performed manually, plays a crucial role in validating machine-extracted metadata. However, manual labeling is time-consuming and not always feasible; thus, there is a need to develop automated text labeling techniques in order to accelerate the process of scientific innovation. This need is particularly urgent in fields such as environmental genomics and microbiome science, which have historically received less attention in terms of metadata curation and creation of gold-standard text mining data sets. In this paper, we present two novel automated text labeling approaches for the validation of ML-generated metadata for unlabeled texts, with specific applications in environmental genomics. Our techniques show the potential of two new ways to leverage existing information that is only available for select documents within a corpus to validate ML models, which can then be used to describe the remaining documents in the corpus. The first technique exploits relationships between different types of data sources related to the same research study, such as publications and proposals. The second technique takes advantage of domain-specific controlled vocabularies or ontologies. In this paper, we detail applying these approaches in the context of environmental genomics research for ML-generated metadata validation. Our results show that the proposed label assignment approaches can generate both generic and highly specific text labels for the unlabeled texts, with up to 44% of the labels matching with those suggested by a ML keyword extraction algorithm.

Crash modification factor (CMF) is an effectiveness measure of safety countermeasures. It is widely used by state agencies to evaluate and prioritize various safety improvement projects. The Federal Highway Administration (FHWA) CMF Clearinghouse provides CMFs for a broad range of countermeasures, but still, the existing CMFs often cannot meet the needs for characterizing the safety impacts of countermeasures in new scenarios. Developing CMFs, meanwhile, is costly, time-consuming, and requires extensive data collection. A more cost-effective way to provide preliminary CMF estimations is needed. To address this need, this study develops a low-cost and easily extendable data-driven framework for CMF predictions. This framework performs data mining on existing CMF records in the FHWA CMF Clearinghouse. To tackle the heterogeneity of data, interdisciplinary techniques to maintain model compatibility were created and used. The project also integrates multiple machine-learning models to learn the complex hidden relationships between different safety countermeasure scenarios. Finally, the proposed framework is trained against the CMF Clearinghouse data and performs comprehensive evaluations. The results show that the proposed framework can provide CMF predictions for new countermeasure scenarios with reasonable accuracy, with overall mean absolute errors less than 0.2. We also discuss an enhanced approach that leverages structured information in certain CMF descriptions, which can boost the CMF prediction accuracy, showing a mean absolute error less than 0.1 in a case study.

Introduction: The emergence of artificial intelligence (AI)-based support system endoscopy, including cystoscopy, has shown promising results by training deep learning algorithms with large datasets of images and videos. This AI-aided cystoscopy has the potential to significantly transform the urological practice by assisting the urologists in identifying malignant areas, especially considering the diverse appearance of these lesions. Methods: Four databases, the PubMed, ProQuest, EBSCOHost, and ScienceDirect were searched, along with a manual hand search. Prospective and retrospective studies, experimental studies, cross-sectional studies, and case-control studies assessing the utilization of AI for the detection of bladder cancer through cystoscopy and comparing with the histopathology results as the reference standard were included. The following terms and their variants were used: 'artificial intelligence,' 'cystoscopy,' and 'bladder cancer.' The risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. A random effects model was used to calculate the pooled sensitivity and specificity. The Moses-Littenberg model was used to derive the Summary Receiver Operating Characteristics (SROC) curve. Results: Five studies were selected for the analysis. Pooled sensitivity and specificity were 0.953 (95 confidence interval [CI]: 0.908-0.976) and 0.957 (95 CI: 0.923-0.977), respectively. Pooled diagnostic odd ratio was 449.79 (95 CI: 12.42-887.17). SROC curve (area under the curve: 0.988, 95 CI: 0.982-0.994) indicated a strong discriminating power of AI-aided cystoscopy in differentiation normal or benign bladder lesions from the malignant ones. Conclusions: Although the utilization of AI for aiding in the detection of bladder cancer through cystoscopy remains questionable, it has shown encouraging potential for enhancing the detection rates. Future studies should concentrate on identification of the patients groups which could derive maximum benefit from accurate identification of the bladder cancer, such as those with intermediate or high-risk invasive tumors., Author(s): Antoninus Hengky (corresponding author) [1,2]; Stevan Kristian Lionardi [3]; Christopher Kusumajaya [4] INTRODUCTION As science and technology continues to advance, the increase in the volume of data, along with [...]

This paper summarizes librarian research on information visualization as well as general trends in the broader field, highlighting the most recent trends, important journals, and which subject disciplines are most involved with information visualization. By comparing librarian research to the broader field, the paper identifies opportunities for libraries to improve their information visualization support services. [ABSTRACT FROM AUTHOR]

Today, colorectal cancer (CRC) diagnosis is performed using colonoscopy, which is the current, most effective screening method. However, colonoscopy poses risks of harm to the patient and is an invasive process. Recent research has proven metabolomics as a potential, non-invasive detection method, which can use identified biomarkers to detect potential cancer in a patients body. The aim of this study is to develop a machine-learning (ML) model based on chemical descriptors that will recognize CRC-associated metabolites. We selected a set of metabolites found as the biomarkers of CRC, confirmed that they participate in cancer-related pathways, and used them for training a machine-learning model for the diagnostics of CRC. Using a set of selective metabolites and random compounds, we developed a range of ML models. The best performing ML model trained on Stage 0-2 CRC metabolite data predicted a metabolite class with 89.55% accuracy. The best performing ML model trained on Stage 3-4 CRC metabolite data predicted a metabolite class with 95.21% accuracy. Lastly, the best-performing ML model trained on Stage 0-4 CRC metabolite data predicted a metabolite class with 93.04% accuracy. These models were then tested on independent datasets, including random and unrelated-disease metabolites. In addition, six pathways related to these CRC metabolites were also distinguished: aminoacyl-tRNA biosynthesis; glyoxylate and dicarboxylate metabolism; glycine, serine, and threonine metabolism; phenylalanine, tyrosine, and tryptophan biosynthesis; arginine biosynthesis; and alanine, aspartate, and glutamate metabolism. Thus, in this research study, we created machine-learning models based on metabolite-related descriptors that may be helpful in developing a non-invasive diagnosis method for CRC.

Classification as an essential part of Machine Learning and Data Mining has significant roles in engineering, medicine, agriculture, military, etc. With the evolution of data collection tools and the unceasing efforts of researchers, new datasets with huge dimensions are obtained so that each data sample has multiple labels. This kind of classification is called Multi-Class Classification (MLC) and demands new techniques to predict the set of labels for a data instance. To date, a variety of methods have been proposed to solve MLC problems. However, new high-dimensional datasets with challenging patterns are being developed, making it necessary for new research to be conducted to develop more efficient methods. This paper presents a novel framework named QLHA to solve MLC problems more efficiently. In the QLHA, the Principal Label Space Transformation (PLST) and Ridge Regression (RR) are recruited to predict the labels of data. Next, an effective objective function is introduced. Also, a hybrid metaheuristic algorithm called QGTOJS is provided to optimize objective value and enhance the predicted labels by selecting the most relevant features. In the QGTOJS, the Gorilla Troops Optimization (GTO) and Jellyfish Search algorithm (JS) are binarized and hybridized through a modified variant of the Q-learning algorithm. Besides, an adjusted Hill Climbing strategy is adopted to balance the exploration and exploitation and improve local optima departure. Likewise, a local search mechanism is provided to enhance searchability as much as possible. Eventually, the QLHA is applied to ten multi-label datasets and the obtained results are compared with heuristic and metaheuristic-based MLC methods numerically and visually. The experimental results disclosed the effectiveness of the contributions and superiority of the QLHA over competitors. [ABSTRACT FROM AUTHOR]

We report evidence of an undocumented method to manipulate citation counts involving "sneaked" references. Sneaked references are registered as metadata for published scientific articles in which they do not appear. This manipulation exploits trusted relationships between various actors: publishers, the Crossref metadata registration agency, digital libraries, and bibliometric platforms. By collecting metadata from various sources, we show that extra undue references are actually sneaked in at Digital Object Identifier (DOI) registration time, resulting in artificially inflated citation counts. As a case study, focusing on three journals from a given publisher, we identified at least 9% sneaked references (5978⁄65,836) mainly benefiting two authors. Despite not being present in the published articles, these sneaked references exist in metadata registries and inappropriately propagate to bibliometric dashboards. Furthermore, we discovered "lost" references: the studied bibliometric platform failed to index at least 56% (36,939/65,836) of the references present in the HTML version of the publications. This research led to an investigation by Crossref (confirming our findings) and to subsequent corrective actions. The extent of the distortion—due to sneaked and lost references—in the global literature remains unknown and requires further investigations. Bibliometric platforms producing citation counts should identify, quantify, and correct these flaws to provide accurate data to their patrons and prevent further citation gaming. [ABSTRACT FROM AUTHOR]

This study employs advanced data mining techniques to investigate the DASS‐42 questionnaire, a widely used psychological assessment tool. Administered to 680 students at Necmettin Erbakan University's Ahmet Kelesoglu Faculty of Education, the DASS‐42 comprises three distinct subscales—depression, anxiety and stress—each consisting of 14 items. Departing from traditional statistical methodologies, the study harnesses the power of the WEKA data mining program to analyse the dataset. Employing Naive Bayes (NB), Artificial Neural Network (ANN), Logistic Regression (LR), Support Vector Machine (SVM) and Random Forest (RF) algorithms, the research unveils novel insights. The ANN method emerges as a standout performer, achieving remarkable distinctiveness scores for all subscales: depression (99.26%), anxiety (98.67%) and stress (97.35%). The study highlights the potential of data mining in enhancing psychological assessment and showcases the ANN's prowess in capturing intricate patterns within complex psychological dimensions. By charting a course beyond conventional statistical methods, this research pioneers a new frontier for employing data mining within the realm of social sciences. As a result of the study, it is recommended that teacher candidates in the teacher education process should have knowledge about depression, anxiety and stress, and relevant courses on these topics should be added to the curriculum of teacher education programs. [ABSTRACT FROM AUTHOR]

Due to the limitation of spraying conditions, the microstructure of Thermal Barrier Coatings (TBCs) is not homogeneous, so it is important to extract the peak information of the reflected signal efficiently and accurately using Terahertz (THz) non-destructive detection technique for online thickness measurement. To this end, a hybrid model peak information extraction method based on a convolutional neural network (CNN) and a gated recurrent unit (GRU) is proposed. Firstly, a theoretical model of the THz signal is used to generate the simulated signal; secondly, a 1D-CNN is used to extract features adaptively from the temporal signal input, and a GRU is constructed to learn the temporal information between feature vectors to complete the extraction of peak information; finally, a calibration strategy is employed to determine the highest value around the original localisation result in order to eliminate localisation error and achieve peak location. The proposed CNN-GRU model predicts evaluation metrics that are 2–4 times smaller than those of other methods, and the running time is reduced to 45.71% compared to networks with close prediction accuracy. The method can reduce manual involvement and time expenses while maintaining prediction accuracy and that it provides a new intelligent way for online thickness measuring. [ABSTRACT FROM AUTHOR]

The pivotal role of dam infrastructure necessitates continuous health monitoring, which results in extensive sets of data. Most monitoring data‐based models in dam engineering concentrate on predicting dam behavior. However, little attention has been systematically paid to the processing of extensive monitoring data, modeling of comprehensive dam behavior, and assessment of overall dam operation status. Here, we propose a novel monitoring model comprising three main aspects: a multidimensional data mining method, a multipoint response prediction method, and a multilayer data fusion‐based assessment method. Utilizing monitoring data from a mega concrete arch dam, we evaluate and discuss the effects of data mining, modeling accuracy for dam behavior, robustness against data pollution, and sensitivity to anomalies. Comparisons with classical benchmarks demonstrate the performance of the proposed model for the dam. [ABSTRACT FROM AUTHOR]

Rapidly emerging generative artificial intelligence (GenAI) models stand at the epicentre of current public discourse. They demonstrate impressive abilities to generate various types of data promptly and cost-effectively. However, AI developers need to train their systems on massive volumes of data which is usually copyrighted. Therefore, the growth of copyright-related concerns in the field of GenAI comes as no surprise. The study introduces two solutions which could mitigate the tension between copyright holders and AI developers, one legal (text and data mining (TDM) exceptions of the CDSM Directive) and one technical (synthetic data), highlighting the promises and challenges of both. First, the article will discuss the capability of TDM exceptions to facilitate the fundamental right to information and the freedom of research in the context of AI development. Next, the paper will analyse how providers of GenAI models can leverage synthetic data to comply with copyright law while training their systems and what risks might be associated with this approach. The findings of this study will indicate what issues, in both legal and technical spheres, should be addressed to ensure a balance of powers in the digital environment and effective functionality of the EU AI sector. [ABSTRACT FROM AUTHOR]

Natural stones are subjected to some processes in stone processing plants for producing stone products, such as cladding, paving, tiling, and decorative materials. Cutting is one of the most critical processes in stone processing. Diamond segments are still widely used in cutting and other processing steps. The parameter that affects the usage of segments is diamond segment wear. Nowadays, unit wear (UW) on the diamond segment should be kept at a minimum in cutting operations for economic production. Thus, the prediction of UW has become a vital issue in stone processing. In this study, UW on diamond segments after stone cutting was evaluated in terms of stone characteristics, operating parameters of the stone cutting machine, vibration amplitude, and sound level measured during cutting. Conventional statistical (CS) methods and data mining (DM) techniques were used to predict unit wear, and these methods were compared. Performed assessments showed that almost all DM techniques give more reliable results than CS methods. Artificial neural networks (ANN) and k-nearest neighbor (k-NN) techniques significantly predicted the UW values more accurately than the other assessment techniques. The results showed a high coefficient of determination with ANN and k-NN obtained as R2= 0.936 and 0.919, respectively. DM techniques can be more efficient in evaluating complex cutting data. [ABSTRACT FROM AUTHOR]

Background: Both learning and problem solving are major goals of complex problem solving in engineering education. The order of knowledge construction and problem solving in learning through problem solving, however, has not been explained in current literature. Objectives: To understand their relationships, this study compared the effects of different goal type orders on collaborative simulation‐based inquiry in engineering problem solving. Methods: In the study, 24 engineering undergraduate students worked in pairs to solve a wind tunnel problem using a computer simulation application. Process mining and epistemic network analysis were used to model the inquiry process and identify inquiry patterns based on the log data and discussion discourse. Results and Conclusions: The results showed the goal type of first assigned subtask influenced the inquiry strategies of medium‐performance learners most. In addition, the goal type of first assigned subtask had influence on learners' inquiry process. The learners assigned with a problem‐solving goal subtask first (Problem solving goal‐Learning goal condition) had more simulation trials and centered their inquiry discourse around clarification while the learners assigned with a learning goal subtask first (Learning goal‐Problem solving goal condition) connected reflection with other inquiry phases more often. The findings of this study suggested assigning a learning goal subtask first in simulated‐based complex engineering problem solving could be used as an idea‐centered scaffolding, especially for medium‐performance learners. However, low‐performance learners would experience a "situate knowledge paradox" and needs extra guidance. It is important to monitor the knowledge construction process in complex problem solving, even when seeking a solution is the major task. Lay Description: What is currently known about this topic: Simulation‐based environment is effective in supporting inquiry problem solving with proper guidance.Goal type is an important instructional element that affects students' achievements and motivation.Dividing a complex problem‐solving task into subtasks is a way to guide the inquiry. What this paper adds: The goal type of first assigned subtask influenced the inquiry strategies of medium‐performance learners the most in simulated inquiry‐based engineering problem solving.Learners assigned with a problem‐solving goal subtask first had more simulation trials and centered their inquiry discourse around clarificationLearners assigned with a learning goal subtask first connected reflection with other inquiry phases more often in their inquiry process. Implications for practice and/or policy: Assigning a learning goal subtask first in simulated‐based complex engineering problem solving could be used as "idea‐centered" scaffolding, especially for medium‐performance learners.Low‐performance learners low‐performance learners would experience a "situate knowledge paradox" and needs extra guidance in complex problem solving.It is important to monitor the knowledge construction process in complex problem solving, even when seeking a solution is the major task. [ABSTRACT FROM AUTHOR]

Background: Computational thinking is derived from arguments that the underlying practices in computer science augment problem‐solving. Most studies investigated computational thinking development as a function of learners' factors, instructional strategies and learning environment. However, the influence of the wider community such as educational policies on computational thinking remains unclear. Objectives: This study examines the impact of basic and technology‐related educational policies on the development of computational thinking. Methods: Using supervised machine learning, the computational thinking achievements of 31,823 eighth graders across nine countries were analysed. Seven rule‐based and tree‐based classification models were generated and triangulated to determine how educational policies predicted students' computational thinking. Results and conclusions: Predictions show that students have a higher propensity to develop computational thinking skills when schools exercise full autonomy in governance and explicitly embed computational thinking in their curriculum. Plans to support students, teachers and schools with technology or introduce 1:1 computing have no discernible predicted influence on students' computational thinking achievement. Implications: Although predictions deduced from these attributes are not generalizable, traces of how educational policies affect computational thinking exist to articulate more fronts for future research on the influence of educational policies on computational thinking. Lay description: What is already known about this topicComputational thinking (CT) is a problem‐solving skill.Inquiries on CT focus on learners' factors such as age, gender and attitudes.Also, the choice of instructional strategies and learning environment influence CT development. What this paper addsArticulated how the wider community structures influence the development of CT.Educational policies affect the development of computational thinking. Implications for practicesStudents have a higher predicted propensity to develop CT when schools exercise full autonomy in governance and embed CT in the curriculum explicitly.Plans to support students, teachers and schools with technology or plans to introduce 1:1 computing have no discernible influence on students' CT. [ABSTRACT FROM AUTHOR]

PURPOSE: Patient-centered outcomes (PCOs) are pivotal in cancer treatment, as they directly reflect patients' quality of life. Although multiple studies suggest that factors affecting breast cancer–related morbidity and survival are influenced by treatment side effects and adherence to long-term treatment, such data are generally only available on a smaller scale or from a single center. The primary challenge with collecting these data is that the outcomes are captured as free text in clinical narratives written by clinicians. MATERIALS AND METHODS: Given the complexity of PCO documentation in these narratives, computerized methods are necessary to unlock the wealth of information buried in unstructured text notes that often document PCOs. Inspired by the success of large language models (LLMs), we examined the adaptability of three LLMs, GPT-2, BioGPT, and PMC-LLaMA, on PCO tasks across three institutions, Mayo Clinic, Emory University Hospital, and Stanford University. We developed an open-source framework for fine-tuning LLM that can directly extract the five different categories of PCO from the clinic notes. RESULTS: We found that these LLMs without fine-tuning (zero-shot) struggle with challenging PCO extraction tasks, displaying almost random performance, even with some task-specific examples (few-shot learning). The performance of our fine-tuned, task-specific models is notably superior compared with their non–fine-tuned LLM models. Moreover, the fine-tuned GPT-2 model has demonstrated a significantly better performance than the other two larger LLMs. CONCLUSION: Our discovery indicates that although LLMs serve as effective general-purpose models for tasks across various domains, they require fine-tuning when applied to the clinician domain. Our proposed approach has the potential to lead more efficient, adaptable models for PCO information extraction, reducing reliance on extensive computational resources while still delivering superior performance for specific tasks. [ABSTRACT FROM AUTHOR]

PURPOSE: The RECIST guidelines provide a standardized approach for evaluating the response of cancer to treatment, allowing for consistent comparison of treatment efficacy across different therapies and patients. However, collecting such information from electronic health records manually can be extremely labor-intensive and time-consuming because of the complexity and volume of clinical notes. The aim of this study is to apply natural language processing (NLP) techniques to automate this process, minimizing manual data collection efforts, and improving the consistency and reliability of the results. METHODS: We proposed a complex, hybrid NLP system that automates the process of extracting, linking, and summarizing anticancer therapy and associated RECIST-like responses from narrative clinical text. The system consists of multiple machine learning–/deep learning–based and rule-based modules for diverse NLP tasks such as named entity recognition, assertion classification, relation extraction, and text normalization, to address different challenges associated with anticancer therapy and response information extraction. We then evaluated the system performances on two independent test sets from different institutions to demonstrate its effectiveness and generalizability. RESULTS: The system used domain-specific language models, BioBERT and BioClinicalBERT, for high-performance therapy mentions identification and RECIST responses extraction and categorization. The best-performing model achieved a 0.66 score in linking therapy and RECIST response mentions, with end-to-end performance peaking at 0.74 after relation normalization, indicating substantial efficacy with room for improvement. CONCLUSION: We developed, implemented, and tested an information extraction system from clinical notes for cancer treatment and efficacy assessment information. We expect this system will support future cancer research, particularly oncologic studies that focus on efficiently assessing the effectiveness and reliability of cancer therapeutics. [ABSTRACT FROM AUTHOR]

PURPOSE: Surgical pathology reports are critical for cancer diagnosis and management. To accurately extract information about tumor characteristics from pathology reports in near real time, we explore the impact of using domain-specific transformer models that understand cancer pathology reports. METHODS: We built a pathology transformer model, Path-BigBird, by using 2.7 million pathology reports from six SEER cancer registries. We then compare different variations of Path-BigBird with two less computationally intensive methods: Hierarchical Self-Attention Network (HiSAN) classification model and an off-the-shelf clinical transformer model (Clinical BigBird). We use five pathology information extraction tasks for evaluation: site, subsite, laterality, histology, and behavior. Model performance is evaluated by using macro and micro F 1 scores. RESULTS: We found that Path-BigBird and Clinical BigBird outperformed the HiSAN in all tasks. Clinical BigBird performed better on the site and laterality tasks. Versions of the Path-BigBird model performed best on the two most difficult tasks: subsite (micro F 1 score of 72.53, macro F 1 score of 35.76) and histology (micro F 1 score of 80.96, macro F 1 score of 37.94). The largest performance gains over the HiSAN model were for histology , for which a Path-BigBird model increased the micro F 1 score by 1.44 points and the macro F 1 score by 3.55 points. Overall, the results suggest that a Path-BigBird model with a vocabulary derived from well-curated and deidentified data is the best-performing model. CONCLUSION: The Path-BigBird pathology transformer model improves automated information extraction from pathology reports. Although Path-BigBird outperforms Clinical BigBird and HiSAN, these less computationally expensive models still have utility when resources are constrained. Exploring advancements in cancer pathology research: Path-BigBird, a specialized transformer model, shows promise in extracting crucial tumor information from SEER cancer registry reports. Outperforming HiSAN and Clinical BigBird, it offers improved insights. A step toward enhanced pathology report analysis. #PathologyResearch #SEER #Transformers #LLM [ABSTRACT FROM AUTHOR]

Following a comprehensive analysis of the initial three generations of prisoner risk assessment tools, the field has observed a notable prominence in the integration of fourth-generation tools and machine learning techniques. However, limited efforts have been made to address the explainability of data-driven prediction models and their connection with treatment recommendations. Our primary objective was to develop predictive models for assessing the likelihood of recidivism among prisoners released from their index incarceration within 1-year, 2-year, and 5-year timeframes. We aimed to enhance interpretability using SHapley Additive exPlanations (SHAP). We collected data from 20,457 in-prison records from February 10, 2005, to August 25, 2021, sourced from a Southwestern China prison's data management system. Recidivism records were officially determined through data mining from an official website and combined identification data from neighboring prisons. We employed five machine learning algorithms, considering sociodemographic, physical health, psychological assessments, criminological characteristics, crime history, social support, and in-prison behaviors as factors. For interpretability, SHAP was applied to reveal feature contributions. Findings indicated that young prisoners accused of larceny, previous convictions, lower fines, and limited family support faced higher reoffending risk. Conversely, middle-aged and senior prisoners with no prior convictions, lower monthly supermarket expenses, and positive psychological test results had lower reoffending risk. We also explored interactions between significant predictive features, such as prisoner age at incarceration initiation and primary accusation, and the duration of current incarceration and cumulative prior incarcerations. Notably, our models consistently exhibited high performance, as shown by AUC on the test dataset across time windows. Interpretability results provided insights into evolving risk factors over time, valuable for intervention with high-risk individuals. These insights, with additional validation, could offer dynamic prisoner information for stakeholders. Moreover, interpretability results can be seamlessly integrated into prison and court management systems as a valuable risk assessment tool. [ABSTRACT FROM AUTHOR]

Nowadays, social networks are inseparable part of people everyday life. Large amounts of check-ins collected by location-based social networks (LBSNs) are rich information sources of social and spatio-temporal aspects of human life. In LBSNs, a co-occurrence refers to an event that a pair (almost) simultaneously appears at the same location. There are very few efforts on co-occurrence prediction in LBSNs in the literature. This paper aims to predict the co-occurrences between pairs of users, whether in the future or the past. This can be used in epidemiological studies like today COVID-19 pandemic. We have identified several effective spatial/temporal/social factors. For each, the method of the predictability measurement has been proposed. The spatial, regional, categorical and hierarchical-temporal aspects have been measured with inverse diversity, which is based on Entropy. In general, lower entropy implies higher predictability. Coincidences are unintentional encounters between users. Unlike previous studies that eliminated them, we have been considered them as a co-occurrence type and their effects have been measured and considered in the prediction task. By utilizing these identified factors along with the friendship information, a novel framework has been developed to enable the co-occurrence pattern inference among user pairs. Furthermore, valid patterns have been identified to prevent predictions based on expired patterns. To the best of our knowledge, this issue has not been studied in any other human behavior investigations. Evaluations with two real LBSN datasets, show the efficacy of our co-occurrence inference framework, outperforming other comparable method. [ABSTRACT FROM AUTHOR]

In general in ICUs, operation theatres, post-operative critical care units, and even ambulatory monitors, the patients are continuously examined with ECG and pulse oximeters PPG. In these situations, where the ECG and/or PPG are afflicted by severe artifacts, the idea of extracting respiratory signal from both ECG and PPG signals rather than from any one of them is tested in this work. As respiratory trend is present in both ECG and PPG signals, the common respiratory trend present in simultaneously recorded ECG and PPG signals is extracted, using a bivariate vector autoregressive modeling (BVAR) technique. This technique effectively reduced the inevitable artifacts and resulted in better estimation of the respiratory activity. For further improving the performance of the BVAR method, in extracting respiratory activity from ECG and PPG signals corrupted with sever artifacts, and also works for broad range of breathing rates, an improved BVAR (IB-VAR) technique is proposed. This technique is robust in the sense that it, firstly, works well even in the presence of various artifacts present in either of the signals, and extracts the signal common to both i.e. respiratory information, with a greater accuracy. Secondly, it also works even for a broad range of breathing rates covering as low as 6 breaths per minute (bpm) to as high as 90 bpm. The novel part of the proposed IB-VAR method is that the respiratory pole lying in that broad breathing range is automatically selected from among all other possible poles, which also include the ones corresponding to noises like motion artifact (MA) and baseline wander (BLW), making use of kurtosis values of extracted signals. An analog front end is developed to record ECG, PPG and respiratory signals with different breathing rates and respiration patterns simultaneously from the volunteers. The method, applied on the recorded data of fifteen healthy subjects, performed extremely well even in the presence of MA and BLW, compared to the well known wavelet based approach. Correlation analysis, done in both frequency and time domains, has shown a high degree of acceptance for the extracted respiratory signal with respect to the original reference respiratory signal. Higher values of accuracy rate (EDR: 98.10 ± 1.45, PDR: 98.45 ± 1.30) and lower values of NRMSE calculations (EDR: − 6.47 ± 4.29, PDR: − 6.50 ± 4.17) clearly confirmed the validity of the extracted respiratory signal. An important finding of this work is that the PPG derived respiratory signal very closely matched with the original than the ECG derived signal. [ABSTRACT FROM AUTHOR]

The paper is devoted to testing Kohonen self‐organising maps, with elliptic Fourier coefficients as quantitative variables, for the task of morphological classification of fragmented and non‐standardised ceramics. The advantage of the methodology used is its ability to account for the systematic statistical relationships inherent in the dataset, build models of varying degrees of generalisation and visualise multivariate data. The approbation of the method was carried out on materials from the Hlyboke Ozero‐2 settlement in Eastern Ukraine. The results are compared with the results obtained using principal component analysis + k‐means clustering. [ABSTRACT FROM AUTHOR]

Research in Natural Language Processing (NLP) is increasing rapidly; as a result, a large number of research papers are being published. It is challenging to find the contributions of the research paper in any specific domain from the huge amount of unstructured data. There is a need for structuring the relevant contributions in Knowledge Graph (KG). In this paper, we describe our work to accomplish four tasks toward building the Scientific Knowledge Graph (SKG). We propose a pipelined system that performs contribution sentence identification, phrase extraction from contribution sentences, Information Units (IUs) classification, and organize phrases into triplets (subject, predicate, object) from the NLP scholarly publications. We develop a multitasking system (ContriSci) for contribution sentence identification with two supporting tasks, viz.Section Identification and Citance Classification. We use the Bidirectional Encoder Representations from Transformers (BERT)—Conditional Random Field (CRF) model for the phrase extraction and train with two additional datasets: SciERC and SciClaim. To classify the contribution sentences into IUs, we use a BERT-based model. For the triplet extraction, we categorize the triplets into five categories and classify the triplets with the BERT-based classifier. Our proposed approach yields the F1 score values of 64.21%, 77.47%, 84.52%, and 62.71% for the contribution sentence identification, phrase extraction, IUs classification, and triplet extraction, respectively, for non-end-to-end setting. The relative improvement for contribution sentence identification, IUs classification, and triplet extraction is 8.08, 2.46, and 2.31 in terms of F1 score for the NLPContributionGraph (NCG) dataset. Our system achieves the best performance (57.54% F1 score) in the end-to-end pipeline with all four sub-tasks combined. We make our codes available at: https://github.com/92Komal/pipeline_triplet_extraction. [ABSTRACT FROM AUTHOR]