Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Myoungkyu Song has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    SYNCode: Synergistic Human–LLM Collaboration for Enhanced Data Annotation in Stack Overflow

    by Meng Xia, Shradha Maharjan, Tammy Le, Will Taylor, Myoungkyu Song

    Large language models (LLMs) have rapidly advanced natural language processing, showcasing remarkable effectiveness as automated annotators across various applications. Despite their potential to significantly reduce annotation costs and expedite workflows, annotations produced solely by LLMs can suffer from inaccuracies and inherent biases, highlighting the necessity of maintaining human oversight. In this article, we present a synergistic human–LLM collaboration approach for data annotation enhancement (SYNCode). This framework is designed explicitly to facilitate collaboration between humans and LLMs for annotating complex, code-centric datasets such as Stack Overflow. The proposed approach involves an integrated pipeline that initially employs TF-IDF analysis for quick identification of relevant textual elements. Subsequently, we leverage advanced transformer-based models, specifically NLP Transformer and UniXcoder, to capture nuanced semantic contexts and code structures, generating more accurate preliminary annotations. Human annotators then engage in iterative refinement, validating and adjusting annotations to enhance accuracy and mitigate biases introduced during automated labeling. To operationalize this synergistic workflow, we developed the SYNCode prototype, featuring an interactive graphical interface that supports real-time collaborative annotation between humans and LLMs. This enables annotators to iteratively refine and validate automated suggestions effectively. Our integrated human–LLM collaborative methodology demonstrates considerable promise in achieving high-quality, reliable annotations, particularly for domain-specific and technically demanding datasets, thereby enhancing downstream tasks in software engineering and natural language processing.


  • Open Access

    Tool Support for Improving Software Quality in Machine Learning Programs

    by Kwok Sun Cheng, Pei-Chi Huang, Tae-Hyuk Ahn, Myoungkyu Song

    Machine learning (ML) techniques discover knowledge from large amounts of data. Modeling in ML is becoming essential to software systems in practice. The accuracy and efficiency of ML models have been focused on ML research communities, while there is less attention on validating the qualities of ML models. Validating ML applications is a challenging and time-consuming process for developers since prediction accuracy heavily relies on generated models. ML applications are written by relatively more data-driven programming based on the black box of ML frameworks. All of the datasets and the ML application need to be individually investigated. Thus, the ML validation tasks take a lot of time and effort. To address this limitation, we present a novel quality validation technique that increases the reliability for ML models and applications, called MLVal . Our approach helps developers inspect the training data and the generated features for the ML model. A data validation technique is important and beneficial to software quality since the quality of the input data affects speed and accuracy for training and inference. Inspired by software debugging/validation for reproducing the potential reported bugs, MLVal takes as input an ML application and its training datasets to build the ML models, helping ML application developers easily reproduce and understand anomalies in the ML application. We have implemented an Eclipse plugin for MLVal that allows developers to validate the prediction behavior of their ML applications, the ML model, and the training data on the Eclipse IDE. In our evaluation, we used 23,500 documents in the bioengineering research domain. We assessed the ability of the MLVal validation technique to effectively help ML application developers: (1) investigate the connection between the produced features and the labels in the training model, and (2) detect errors early to secure the quality of models from better data. Our approach reduces the cost of engineering efforts to validate problems, improving data-centric workflows of the ML application development.


  • Open Access

    MegaD: Deep Learning for Rapid and Accurate Disease Status Prediction of Metagenomic Samples

    by Yassin Mreyoud, Myoungkyu Song, Jihun Lim, Tae-Hyuk Ahn

    The diversity within different microbiome communities that drive biogeochemical processes influences many different phenotypes. Analyses of these communities and their diversity by countless microbiome projects have revealed an important role of metagenomics in understanding the complex relation between microbes and their environments. This relationship can be understood in the context of microbiome composition of specific known environments. These compositions can then be used as a template for predicting the status of similar environments. Machine learning has been applied as a key component to this predictive task. Several analysis tools have already been published utilizing machine learning methods for metagenomic analysis. Despite the previously proposed machine learning models, the performance of deep neural networks is still under-researched. Given the nature of metagenomic data, deep neural networks could provide a strong boost to growth in the prediction accuracy in metagenomic analysis applications. To meet this urgent demand, we present a deep learning based tool that utilizes a deep neural network implementation for phenotypic prediction of unknown metagenomic samples. (1) First, our tool takes as input taxonomic profiles from 16S or WGS sequencing data. (2) Second, given the samples, our tool builds a model based on a deep neural network by computing multi-level classification. (3) Lastly, given the model, our tool classifies an unknown sample with its unlabeled taxonomic profile. In the benchmark experiments, we deduced that an analysis method facilitating a deep neural network such as our tool can show promising results in increasing the prediction accuracy on several samples compared to other machine learning models.


  • Learning To Rank Relevant Documents for Information Retrieval in Bioengineering Text Corpora

    by Kowk Sun Cheng, Myoungkyu Song

    In this paper, we present a Learning To Rank-based approach that helps EXPLORE and understand Relevant documents for bioengineering text corpora, called LTREXPLORER. Based on the likelihood of being the most relevance to a search query, the ranking model sorts documents according to their degrees of relevance, preference, or importance with various domain-specific features. The evaluation results demonstrated that our approach has the potential to effectively provide the retrieval scoring functions to researchers, who focus on the most relevant documents in bioengineering information retrieval.


  • Analyzing Bug Reports by Topic Mining in Software Evolution

    by Uy Nguyen, Kowk Sun Cheng, Samuel Sungmin Cho, Myoungkyu Song

    Reporting bugs is one of the vital activities for evolving software systems. Given such reports, developers cope with unanticipated behaviors during software development, maintenance, and operations. The description of bug reports typically includes (1) what errors occurred previously and (2) how a failure can be reproduced through specific steps, test inputs, and original configurations when a failure was created. However, analyzing bug reports is a tedious and error-prone process due to overflowing, complex terminologies. For example, diverse terms are used to represent similar or divergent elucidations by surrounding contexts during software development and maintenance. To address this problem, we present an approach that applies a topic mining technique to bug reports for finding an adequate code reviewer, who can potentially cope with reported failures, by inferring some hidden topics of a textual document.


  • Open Access

    Using Apache Spark on genome assembly for scalable overlap-graph reduction

    by Alexander J. Paul, Dylan Lawrence, Myoungkyu Song, Seung-Hwan Lim, Chongle Pan, Tae-Hyuk Ahn

    Background: De novo genome assembly is a technique that builds the genome of a specimen using overlaps of genomic fragments without additional work with reference sequence. Sequence fragments (called reads) are assembled as contigs and scaffolds by the overlaps. The quality of the de novo assembly depends on the length and continuity of the assembly. To enable faster and more accurate assembly of species, existing sequencing techniques have been proposed, for example, high-throughput next-generation sequencing and long-reads-producing third-generation sequencing. However, these techniques require a large amounts of computer memory when very huge-size overlap graphs are resolved. Also, it is challenging for parallel computation.Results: To address the limitations, we propose an innovative algorithmic approach, calledScalableOverlap-graphReductionAlgorithms (SORA). SORA is an algorithm package that performs string graph reduction algorithms by Apache Spark. The SORA’s implementations are designed to execute de novo genome assembly on either a single machine or a distributed computing platform. SORA efficiently compacts the number of edges on enormous graphing paths by adapting scalable features of graph processing libraries provided by Apache Spark, GraphX and GraphFrames.Conclusions: We shared the algorithms and the experimental results at our project website,https://github.com/BioHPC/SORA. We evaluated SORA with the human genome samples. First, it processed a nearly one billion edge graph on a distributed cloud cluster. Second, it processed mid-to-small size graphs on a single workstation within a short time frame. Overall, SORA achieved the linear-scaling simulations for the increased computing instances.


  • Open Access

    Tool support for managing repetitive program changes in evolving software

    by Vamshi Krishna Epuri, Sushma Sakala, Tae-Hyuk Ahn, Myoungkyu Song

    Software modification often requires consistent program changes , a group of similar, related changes, at multiple locations in a program. Developers are typically uneasy to (i) detect potential change anomalies such as omission errors or incorrect edits and (ii) determine related locations to apply consistent changes, which is a tedious and error‐prone process. To address this problem, this study presents a technique for managing consistent program changes, checking and applying repetitive program transformation (CARP). Given program differencing results between original and edited program versions, CARP (i) infers change patterns to detect change anomalies , (ii) identifies required edit locations , and (iii) automatically applies adequate edits . It has been implemented in the context of the integrated development environment as an Eclipse plug‐in. The authors evaluated CARP on three open‐source projects and found that CARP detects seeded anomalies with 99.1% accuracy on average. Furthermore, it identifies change locations and transforms them with 98% accuracy. Their results show that CARP should help developers detect potential change anomalies in repetitive program changes and perform consistent changes automatically.