Programming Language Code Research Articles

Analysis of Datasets and Large Language Models for Vulnerability Detection in Imperative Programming Language Code

Large language models are machine learning models that enable the classification and generation of both natural language text and code in various programming languages. These models have billions of parameters and are trained on vast datasets. In recent years, such models have been successfully applied to a wide range of tasks in software engineering. The paper presents data on publication activity on the topic under study, which is obtained on the basis of statistical analysis of search results for relevant key queries. In addition, a review of recent publications in the field of using large language models to detect vulnerabilities in program code is carried out, and the results of the analysis of data sets that are used in training neural network models to search for vulnerabilities in program code are presented.

May the Source Be with You: On ChatGPT, Cybersecurity, and Secure Coding

Software security is an important topic that is gaining more and more attention due to the rising number of publicly known cybersecurity incidents. Previous research has shown that one way to address software security is by means of a serious game, the CyberSecurity Challenges, which are designed to raise awareness of software developers of secure coding guidelines. This game, proven to be very successful in the industry, makes use of an artificial intelligence technique (laddering technique) to implement a chatbot for human–machine interaction. Recent advances in machine learning have led to a breakthrough, with the implementation and release of large language models, now freely available to the public. Such models are trained on a large amount of data and are capable of analyzing and interpreting not only natural language but also source code in different programming languages. With the advent of ChatGPT, and previous state-of-the-art research in secure software development, a natural question arises: to what extent can ChatGPT aid software developers in writing secure software? In this work, we draw on our experience in the industry, and also on extensive previous work to analyze and reflect on how to use ChatGPT to aid secure software development. Towards this, we conduct two experiments with large language models. Our engagements with ChatGPT and our experience in the field allow us to draw conclusions on the advantages, disadvantages, and limitations of the usage of this new technology.

Perceived impact of generative AI on assessments: Comparing educator and student perspectives in Australia, Cyprus, and the United States

The growing use of generative AI tools built on large language models (LLMs) calls the sustainability of traditional assessment practices into question. Tools like OpenAI's ChatGPT can generate eloquent essays on any topic and in any language, write code in various programming languages, and ace most standardized tests, all within seconds. We conducted an international survey of educators and students in higher education to understand and compare their perspectives on the impact of generative AI across various assessment scenarios, building on an established framework for examining the quality of online assessments along six dimensions. Across three universities, 680 students and 87 educators, who moderately use generative AI, consider essay and coding assessments to be most impacted. Educators strongly prefer assessments that are adapted to assume the use of AI and encourage critical thinking, while students' reactions are mixed, in part due to concerns about a loss of creativity. The findings show the importance of engaging educators and students in assessment reform efforts to focus on the process of learning over its outputs, alongside higher-order thinking and authentic applications.

Improving the Community Services through Electronic Management of East Pringsewu Subdistrict Administration

Every Government Institution cannot be separated from carrying out daily administrative activities. Manuscripts or recordings of documents or information, including text, images,and sound recordings, are called archives because archives are records of every activity carried out. Currently, the East Pringsewu sub-district office for filing letters and other activities still uses manual methods. The implementation of E-Government is expected to enable all service activities to the community to be carried out electronically, thereby simplifying policy and service functions. The data collection methods used in this writing are observation, documentation, interviews, and literature study. The SDLC (System Development Life Cycle) method is used in this research, namely a systematic approach to planning, designing, developing, testing,and maintaining software systems. Based on needs analysis, the system developed focuses on managing population data, archiving,and correspondence related to village authority. Before implementing the system, it is necessary to design the system. When designing a system, use the Unified Modeling Language (UML) approach using Use Case Diagrams, Activity Diagrams, and Class Diagrams. Use case diagrams to describe the interaction between users and the system being developed. This research uses the PHP programming language and Visual Studio Code as a text editor and MySQL DBMS. Applications can be used as a medium for implementing subdistrict or village development with information technology and supporting the progress of subdistricts or villages byGovernment recommendations in developing E-Government.

Query-oriented two-stage attention-based model for code search

Applying code search models to search through a large-scale codebase can significantly contribute to developers finding and reusing existing code. Researchers have applied deep learning (DL) techniques to code search models, which first compute deeper semantics representation for query and candidate code snippets, and then rank code snippets. However, these models do not well deeply analyze the semantics gap (i.e., the difference and correlation between queries written in natural language and code in programming languages), or suitably apply the correlation to the code search task. Moreover, most DL-based models use complex networks, slowing down code search tasks.To build the correlation of two languages, and apply the correlation well to code search task, we propose a query-oriented code search model named QobCS. QobCS leverage two attention-based stages, which are simple and quick, and the cooperation of the two stages bridges the semantic gap between code and query. Stage1 learns deeper semantics representation for code and query. Stage2 applies their deeper semantic correlation and query’s intention to learn better code representation.We evaluated QobCS on two datasets. On dataset1/dataset2 with 485k/542k code snippets, QobCS achieves the MRRs of 0.701/0.595, outperforming DL-based code search models DeepCS, CARLCS-CNN, UNIF, and our prior study TabCS. For efficiency, our model shows desirable performances on both datasets compared to DL-based models.

Assessing the Effect of Programming Language and Task Type On Eye Movements of Computer Science Students

Background and Context: Understanding how a student programmer solves different task types in different programming languages is essential to understanding how we can further improve teaching tools to support students to be industry-ready when they graduate. It also provides insight into students’ thought processes in different task types and languages. Few (if any) studies investigate whether any differences exist between the reading and navigation behavior while completing different types of tasks in different programming languages. Objectives: We investigate whether the use of a certain programming language (C++ vs. Python) and type of task (new feature vs. bug fixing) has an impact on performance and eye movement behavior in students exposed to both languages and task types. Participants: Fourteen students were recruited from a Python course that taught Python as an introductory programming language. Study Method: An eye tracker was used to track how student programmers navigate and view source code in different programming languages for different types of tasks. The students worked in the Geany IDE (used also in their course) while eye tracking data was collected behind the scenes making their working environment realistic compared to prior studies. Each task type had a Python and C++ version, albeit on different problems to avoid learning effects. Standard eye tracking metrics of fixation count and fixation durations were calculated on various areas of the screen and on source code lines. Normalized versions of these metrics were used to compare across languages and tasks. Findings: We found that the participants had significantly longer average fixation duration and total fixation duration adjusted for source code length during bug fixing tasks than the feature addition tasks, indicating bug fixing is harder. Furthermore, participants looked at lines adjacent to the line containing the bug more often before looking at the buggy line itself. Participants who added a new feature correctly made their first edit earlier compared to those who failed to add the feature. Tasks in Python and C++ have similar overall fixation duration and counts when adjusted for character count. The participants spent more time fixating on the console output while doing Python tasks. Overall, task type has a bigger effect on the overall fixation duration and count compared to the programming language. Conclusions: CS educators can better support students in debugging their code if they know what they typically look at while bug fixing. For new feature tasks, training students not to fear edits to learn about the code could also be actively taught and encouraged in the classroom. CS education researchers can benefit by building better IDE plugins and tools based on eye movements that guide novices in recognizing bugs and aid in adding features. These results will lead to updating prior theories on mental models in program comprehension of how developers read and understand source code. They will eventually help in designing better programming languages and better methods of teaching programming based on evidence on how developers use them.

Using Natural Language Processing for Programming Language Code Classification with Multinomial Naive Bayes

Using Natural Language Processing for Programming Language Code Classification with Multinomial Naive Bayes

P-096 Prediction of fertilization success by voltage-dependent intracellular Ca2+ dynamics in non-normozoospermic human semen samples undergoing in vitro fertilization: preliminary results

Abstract Study question Does the change in intracellular calcium levels evoked by artificial membrane depolarizations predict fertilization success in a non-normozoospermic semen sample undergoing assisted reproductive techniques? Summary answer The increase in the intracellular calcium after artificial plasma membrane depolarization correlates negatively with fertilization rates in conventional IVF but not with ICSI What is known already Sperm capacitation-related features, like the regulation of the intracellular calcium concentration ([Ca2+]i) and the hyperpolarization of the plasma membrane potential (Vm), have been proposed as predictors of success during artificial reproduction techniques (ART). Intriguingly, upon Vm depolarization, sperm exhibit a rapid and transitory increase in the [Ca2+]i controlled, likely, by the activation of the sperm-specific calcium channel CatSper. The regulation of CatSper activity is essential for sperm function and is crucial for fertilization. Therefore we wondered if the Vm depolarization-dependent rise in the [Ca2+]i can be used to predict fertilization success rates in non-normozoospermic semen samples of men undergoing ART. Study design, size, duration This prospective study employed unidentified semen samples from 29 non-normozoospermic men (37.8 ± 5.6 years old) undergoing fertility treatment from March 2022 to the present at CITMER Reproductive Medicine, Mexico City. Frozen semen samples and procedures with less than 3 oocytes were excluded. Ten normozoospermic semen samples were analyzed as a control. The depolarization-dependent [Ca2+]i changes were evaluated by flow cytometry under four different Vm conditions: resting (variable among samples), –64, –46, and –32 mV. Participants/materials, setting, methods Semen samples were obtained on the same day as oocyte retrieval. Sperm cells were washed and incubated in capacitating HTF media for further use in conventional IVF or ICSI procedures. When available, the surplus cells were stained with vitality, Ca2+, and Vm-sensitive fluorescent dyes. Fluorescence changes were evaluated by AccuriC6 flow cytometer. The kinetical analysis of the transitory [Ca2+]i increase as well as the statistical comparisons were performed using custom-made R programming language code. Main results and the role of chance Analysis of resting Vm showed that normozoospermic semen samples presented more depolarized Vm than non-normozoospermic controls (–75.7 ± 6.9 vs –67.3 ± 10.5 mV, respectively, p = 0.02). The former group of semen samples was either used for conventional IVF (n = 11) or ICSI (n = 14). The semen samples in these last categories showed a tendency to present a more depolarized Vm than non-normozoospermic samples (–69.9 ± 8.8 and –66.6 ± 11.3 mV, respectively). No differences were seen in the resting Ca2+ levels. We then induced a Vm hyperpolarization by adding the K+ ionophore, valinomycin followed by increasing concentrations of extracellular K+. Each depolarizing stimulus evoked a fast and transitory increase in the [Ca2+]i. We then calculated seven different kinetic parameters of such responses: basal Ca2+ level, increase and decrease velocity, maximum Ca2+ reached, response durability, and area under the curve (AUC). The basal Ca2+ level reached after depolarizing stimuli correlated negatively (R = –0.64, p = 0.03) with the conventional IVF fertilization ratio (number of zygotes/total number of mature oocytes) when the Vm was clamped either to –64 and –46 mV. A negative and significant correlation (R = –0.76, p = 0.006) was seen for the AUC analysis at –64 mV. No relevant correlations were found in the ICSI samples. Limitations, reasons for caution This is an in vitro study, and caution must be taken when extrapolating these results in vivo. The preliminary nature of the results implies a necessity to increase the number of analyzed semen samples (in progress). Wider implications of the findings Further research into the molecular regulation of the CatSper channel could offer promising alternatives for clinicians to appropriately choose between IVF and ICSI for each couple. This protocol is fast and simple to perform in a laboratory equipped with a flow cytometer or any fluorescence-based reader. Trial registration number NA

P-094 Dynamics of progesterone-induced intracellular Ca2+ changes in sperm predicts zygote formation rate by IVF: Preliminary results

Abstract Study question Can the overall intracellular Ca2+ dynamics evoked by progesterone stimulation predict the fertilizing potential of a non-normozoospermic semen sample under fertility treatment? Summary answer Some kinetic parameters of the transient intracellular Ca2+ increase evoked by progesterone stimulation correlate with fertilization rates in conventional IVF but not with ICSI. What is known already Stimulation with female hormone progesterone induces the activation of the sperm-specific ion channel CatSper which in turn produces a fast increase in the intracellular Ca2+ concentration ([Ca2+]i). After reaching a maximum [Ca2+]i, different Ca2+ clearance mechanisms become active producing a transient followed by a plateau. We investigated whether the [Ca2+]i dynamics upon progesterone stimulus may be used to assess the fertilizing potential of sperm samples from men undergoing assisted reproduction. Study design, size, duration This prospective study employed unidentified semen samples from 39 non-normozoospermic men (age 38 ± 6 years old) undergoing fertility treatment from March 2022 to the present at CITMER Reproductive Medicine, Mexico City. Frozen semen samples and ART procedures with less than 3 oocytes were excluded. Ten normozoospermic semen samples were also analyzed. The progesterone-dependent [Ca2+]i changes were evaluated using time-lapse flow cytometry. Participants/materials, setting, methods Semen samples were obtained on the same day as oocyte retrieval. Sperm cells were washed and incubated in capacitating HTF media for further use in conventional IVF or ICSI procedures. When available, the surplus cells were stained with vitality and Ca2+-sensitive fluorescent dyes. Fluorescence changes were evaluated by AccuriC6 flow cytometer. Progesterone-induced [Ca2+]i transient as well as the correlations with the fertilization rates were calculated using custom-made R programming language code. Main results and the role of chance The progesterone (1 µM) stimulation evoked a fast and transitory increase in the [Ca2+]i. We then calculated six different kinetics parameters of such responses: basal Ca2+ level, increase and decrease in velocity, maximum Ca2+ level, total response duration, and the area under the curve (AUC). There were no significant differences in any of these parameters between normozoospermic donors (and non-normozoospermic patients. The former group of semen samples were divided into either conventional IVF (n = 17) or ICSI (n = 21). In IVF but not ICSI samples, the duration, and the area under the curve of the progesterone response tended to be higher than in normozoospermic (2.39±0.7 min, p = 0.03, and 113±30 AUF*min, p = 0.18, respectively). Furthermore, both parameters showed a tendency to correlate negatively with the conventional IVF (R= –0.45, p = 0.07 and R= –0. 5, p = 0.04, respectively) but not with ICSI fertilization ratio (number of zygotes/total number of mature oocytes). These results suggest that cells with a better Ca2+ recovery mechanism are well suited for IVF. Interestingly, we observed no difference between the magnitude of the [Ca2+]i increase after progesterone stimulus among donors, IVF, or ICSI samples, as was previously reported. Differences in the concentration of progesterone could explain such results. Limitations, reasons for caution The preliminary nature of the results implies a necessity to increase the number of analyzed semen samples (in progress). It is well known that progesterone induces variable responses even among the same donor cells. Our population-based analysis may be underestimating such single cells variations. Wider implications of the findings Current protocols for sperm analysis fail to successfully predict the fertilizing capability of semen samples from men seeking reproductive treatment. Further research into the molecular regulation of Ca2+ dynamics could offer promising alternatives for clinicians to appropriately choose between IVF and ICSI for each couple. Trial registration number .

PERBANDINGAN ALGORITMA EFFICIENTNETB0 DAN INCEPTIONV3 DALAM KLASIFIKASI CITRA JENIS ANJING

Dogs are pets that have hundreds of breeds. The various breeds of dogs can be distinguished by size, fur color, and type of fur. The identification of dog breeds can use images of dogs as one of the object recognition information. Image classification can be made using Deep Learning methods. Image classification is the process of grouping pixels in an image into a number of classes, so that each class can represent an entity with certain characteristics. Additionally, Deep Learning has many algorithms that can be used for image classification, including the InceptionV3 and EfficientNetB0 algorithms. The purpose of this research is to compare the accuracy results of the EfficientNetB0 and InceptionV3 algorithms in classifying dog breed images. The stages in making this program include collecting datasets, data preparation, model design, and interpreting the results. This image classification is made using the Python programming language and Visual Studio Code software. The program has an accuracy result with the best model being InceptionV3, with an accuracy of 97.30% on the training set and 97.07% on the validation set after 20 epochs.

Python code smells detection using conventional machine learning models.

Code smells are poor code design or implementation that affect the code maintenance process and reduce the software quality. Therefore, code smell detection is important in software building. Recent studies utilized machine learning algorithms for code smell detection. However, most of these studies focused on code smell detection using Java programming language code smell datasets. This article proposes a Python code smell dataset for Large Class and Long Method code smells. The built dataset contains 1,000 samples for each code smell, with 18 features extracted from the source code. Furthermore, we investigated the detection performance of six machine learning models as baselines in Python code smells detection. The baselines were evaluated based on Accuracy and Matthews correlation coefficient (MCC) measures. Results indicate the superiority of Random Forest ensemble in Python Large Class code smell detection by achieving the highest detection performance of 0.77 MCC rate, while decision tree was the best performing model in Python Long Method code smell detection by achieving the highest MCC Rate of 0.89.

Perancangan Sistem Akuntansi Penggajian ‘Si Warna’ Di Pt Harapan Kurnia Textile Indonesia

The research conducted at PT Harapan Kurnia Textile Indonesia aims to produce accurate attendance information, fast, effective and efficient payroll management, provide financial reports for companies and payslips for employees appropriately. To identify the problems faced, primary and secondary data collection as well as qualitative descriptive approaches were used in this study. The method for developing software that uses waterfall with HTML to create pages on websites and databases is MYSQL. The results of this study show that the payroll management process has run well but still uses Microsoft Excel manually which is prone to errors. So to overcome these problems, an application design was made in the form of Flowmap, ERD and DFD and designed a payroll accounting system with web-based PHP programming language code. The conclusion of this study is that the color becomes a problem solver that will help HRD and financial staff in managing payroll to be faster, more precise and efficient and produce accurate payslips and financial statements. The advice given by the author is to conduct computer training payroll applications, perform system maintenance to prevent damage / errors in the system and back up data so that data remains safe if the computer is damaged.
 Keywords : Design, Accounting System, Payroll

An empirical examination of newcomer contribution costs in established OSS communities: a knowledge-based perspective

PurposeTo remain sustainable, open source software (OSS) projects must attract new members—or newcomers—who make contributions. In this paper, the authors develop a set of hypotheses based on the knowledge barriers framework that examines how OSS communities can encourage contributions from newcomers.Design/methodology/approachEmploying longitudinal data from the source code repositories of 232 OSS projects over a two-year period, the authors employ a Poisson-based mixed model to test how community characteristics, such as the main drivers of knowledge-based costs, relate to newcomers' contributions.FindingsThe results indicate that community characteristics, such as programming language choice, documentation effort and code structure instability, are the main drivers of knowledge-based contribution costs. The findings also suggest that managing these costs can result in more inclusive OSS communities, as evidenced by the number of contributing newcomers; the authors highlight the importance of maintaining documentation efforts for OSS communities.Originality/valueThis paper assumes that motivational factors are a necessary but insufficient condition for newcomer participation in OSS projects and that the cost to participation should be considered. Using the knowledge barriers framework, this paper identifies the main knowledge-based costs that hinder newcomer participation. To the best of the authors' knowledge, this is the first empirical study that does not limit data collection to a single hosting platform (e.g., SourceForge), which improves the generalizability of the findings.

Applying the SCRUM Framework in Managing the Development of Candidate Profiling Project (A Case Study of Bagidata)

Bagidata is a people analytics platform that provides various data analysis solutions for Human Resource Management. Bagidata develops and maintains an application for creating candidate profiles using SCRUM and Agile. However, there are still many obstacles in the SCRUM and Agile processes in developing this application, such as all documentation being done manually with Microsoft Excel and separated, the inability to monitor the process, the and lack of an integrated platform for team members to share work and make documentation. Therefore, a system is needed to facilitate application development and become an integrated platform for application work, monitoring, and web-based documentation. This study uses object-oriented methodology or OOAD (Object Oriented Analysis and Design), and the data collection method is through observation, interviews, and literature review. This web-based application uses Javascript programming language, PostgreSQL database, and Visual Studio Code as tools, as well as UML modeling with 4 design diagrams, namely a use case diagram, activity diagram, sequence diagram, and class diagram. The results of this study are expected to enable the product owner and development team to monitor the development process, organize task allocation among team developers, and provide an integrated platform for documentation, thus improving the candidate profiling application development process.

Accounting for Context in Randomized Trials after Assignment

Many preventive trials randomize individuals to intervention condition which is then delivered in a group setting. Other trials randomize higher levels, say organizations, and then use learning collaboratives comprised of multiple organizations to support improved implementation or sustainment. Other trials randomize or expand existing social networks and use key opinion leaders to deliver interventions through these networks. We use the term contextually driven to refer generally to such trials (traditionally referred to as clustering, where groups are formed either pre-randomization or post-randomization — i.e., a cluster-randomized trial), as these groupings or networks provide fixed or time-varying contexts that matter both theoretically and practically in the delivery of interventions. While such contextually driven trials can provide efficient and effective ways to deliver and evaluate prevention programs, they all require analytical procedures that take appropriate account of non-independence, something not always appreciated. Published analyses of many prevention trials have failed to take this into account. We discuss different types of contextually driven designs and then show that even small amounts of non-independence can inflate actual Type I error rates. This inflation leads to rejecting the null hypotheses too often, and erroneously leading us to conclude that there are significant differences between interventions when they do not exist. We describe a procedure to account for non-independence in the important case of a two-arm trial that randomizes units of individuals or organizations in both arms and then provides the active treatment in one arm through groups formed after assignment. We provide sample code in multiple programming languages to guide the analyst, distinguish diverse contextually driven designs, and summarize implications for multiple audiences.

Preparing First-Year Engineering Students to Think About Code: A Guided Inquiry Approach

In the wake of the so-called fourth industrial revolution, computer programming has become a foundational competency across engineering disciplines. Yet engineering students often resist the notion that computer programming is a skill relevant to their future profession. Here are presented two activities aimed at supporting the early development of engineering students’ attitudes and abilities regarding programming in a first-year engineering course. Both activities offer students insights into the way programs are constructed, which have been identified as a source of confusion that may negatively affect acceptance. In the first activity, a structured, language-independent way to approach programming problems through guided questions was introduced, which has previously been used successfully in introductory computer science courses. The team hypothesized that guiding students through a structured reflection on how they construct programs for their class assignments might help reveal an understandable structure to them. Results showed that students in the intervention group scored nearly a full letter grade higher on the unit’s final programming assessment than those in the control condition. The second activity aimed to help students recognize how their experience with MATLAB might help them interpret code in other programming languages. In the intervention group, students were asked to review and provide comments for code written in a variety of programming languages. A qualitative analysis of their reflections examined what skills students reported they used and, specifically, how prior MATLAB experience may have aided their ability to read and comment on the unfamiliar code. Overall, the ability to understand and recognize syntactic constructs was an essential skill in making sense of code written in unfamiliar programming languages. Syntactic constructs, lexical elements, and patterns were all recognized as essential landmarks used by students interpreting code they did not write, especially in new languages. Developing an understanding of the static structure and dynamic flow required of programs was also an essential skill which helped the students. Together, the results from the first activity and the insights gained from the second activity suggest that guided questions to build skills in reading code may help mitigate confusion about program construction, thereby better preparing engineering students for computing-intensive careers.

Thermodynamic analysis of gas turbine performance using the enthalpy–entropy approach

The enthalpy–entropy approach, which is a rigorous and accurate method, was used to model the actual cycle of a gas turbine (GT). The constructed GT C# programming language code was used to study the effects of ambient temperature ( T am ), relative humidity (RH), and ambient pressure ( P am ) on GT output parameters. The results of the GT modeling program showed that the maximum (minimum) difference in GT output power ( P GT ) was 1.12% (0.01%) at a T am of 39.7 °C (23.3 °C) compared with the actual output power of El Atf power station. As T am increased from 14 °C to 40 °C, the P GT decreased by 19.01% (18.38%) of nominal GT power (250 MW) at 15% (95%) RH while GT thermal efficiency ( η th ) decreased by 2.3058% (2.4431%) and the specific fuel consumption (SFC) increased by 7.12% (7.59%). The P GT decreased by 0.1497% and 0.7674% with an RH increase from 15% to 95%, respectively. The P GT decreased by 0.03675 and 0.16835 MW while η th decreased by 0.0047% and 0.0243% at T am values of 15 and 40 °C, respectively, when P am was decreased from 1.013 to 0.990 bar. Using the compressor assumptions of each stage increases the humid air pressure by the same value and the efficiency of each stage is the same and equal to the compressor overall efficiency is more logical and leads to accurate results. The compressor isentropic efficiency was not affected by the bleeding air percent according to the assumptions used.

UCLAONT: Ontology-Based UML Class Models Verification Tool

The software design model performs an important role in modern software engineering methods. Especially in Model-Driven Engineering (MDE), it is treated as an essential asset of software development; even programming language code is produced by the models. If the model has errors, then they can propagate into the code. Model verification tools check the presence of errors in the model. This paper shows how a UML class model verification tool has been built to support complex models and unsupported elements such as XOR constraints and dependency relationships. This tool uses ontology for verifying the UML class model. It takes a class model in XMI format and generates the OWL file. Performs verification of model in two steps: (1) uses the ontology-based algorithm to verify association multiplicity constraints; and (2) uses ontology reasoner for the verification of XOR constraints and dependency relationships. The results show the proposed tool improves the verification efficiency and supports the verification of UML class model elements that have not been supported by any existing tool.

Characteristics of Online Crowdfunding Campaigns for Urological Cancers in the United States.

We examined the characteristics and financial outcomes of online crowdfunding campaigns for patients with major urological cancers in the U.S. This cross-sectional study analyzed publicly available data from GoFundMe, the largest online medical crowdfunding service, via automated web scraping. Online campaigns from 2010 to 2018 with the following primary cancer types were included: kidney, prostate, bladder and testicular. Financial outcomes were compared using Kruskal-Wallis and Wilcoxon rank-sum tests. Multivariable analyses were utilized to identify predictors of campaign financial outcomes. Kidney cancers were the most frequent online campaign type (478), followed by prostate (379), bladder (202) and testicular (175) malignancies. Urological cancer campaign recipients frequently requested funding for medical expenses (71%) during active treatment (57%). After adjustment, testicular cancer and children's cancer campaigns generated more donations than other urological and adult cancer campaigns (p <0.05). Family and friend-authored campaigns generated more donations and average donation amounts than self-authored campaigns (p <0.05). Campaign narratives focused on disheartening circumstances received fewer donations than narratives focused on the recipient's high moral character or contributions to society (p <0.05), and unclear narratives received the smallest donation amounts (p <0.05). Urological cancer crowdfunding in the U.S. is primarily used to finance uncovered costs associated with medical care during active treatment. Crowdfunding financial outcomes are likely related to the campaign recipient's age, malignancy type, social network and primary appeal of the narrative. Urologists should be aware of trends in medical crowdfunding in order to better understand the financial burden this patient population faces.

Deep Graph Matching and Searching for Semantic Code Retrieval

Code retrieval is to find the code snippet from a large corpus of source code repositories that highly matches the query of natural language description. Recent work mainly uses natural language processing techniques to process both query texts (i.e., human natural language) and code snippets (i.e., machine programming language), however, neglecting the deep structured features of query texts and source codes, both of which contain rich semantic information. In this article, we propose an end-to-end deep graph matching and searching (DGMS) model based on graph neural networks for the task of semantic code retrieval. To this end, we first represent both natural language query texts and programming language code snippets with the unified graph-structured data, and then use the proposed graph matching and searching model to retrieve the best matching code snippet. In particular, DGMS not only captures more structural information for individual query texts or code snippets, but also learns the fine-grained similarity between them by cross-attention based semantic matching operations. We evaluate the proposed DGMS model on two public code retrieval datasets with two representative programming languages (i.e., Java and Python). Experiment results demonstrate that DGMS significantly outperforms state-of-the-art baseline models by a large margin on both datasets. Moreover, our extensive ablation studies systematically investigate and illustrate the impact of each part of DGMS.