See a map of all the places I've given a talk!
Conference proceedings talk, 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA), Seattle, Washington, United States
In this talk, I discussed our paper “Enhancing REST API Testing with NLP Techniques”.
Poster, 2023 Georgia Tech PhD Visit Day, Georgia Tech, Georgia, United States
At the Georgia Tech PhD Visit Day, I had the opportunity to present a poster session that showcased our research project titled “Enhancing REST API Testing with NLP Techniques”. This event provided a platform to share the innovative approaches our team has been developing to improve REST API testing, leveraging Natural Language Processing techniques. Our work aims to bridge the gap between human-readable specifications and automated testing tools, leading to significant improvements in the efficiency and accuracy of REST API testing. This poster session was an excellent opportunity to engage with prospective Ph.D. students, sharing insights into the exciting research opportunities available at Georgia Tech.
Conference proceedings talk, 31st ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA), Virtual Conference
In this talk, I discussed our paper “Automated Test Generation for REST APIs: No Time to Rest Yet”. I presented our research on the challenges and limitations of current testing approaches for REST APIs and introduced our novel approach to enhance testing efficiency. Our method leverages machine learning and natural language processing techniques to interpret OpenAPI rules from the specifications’ human-readable part, significantly improving the performance of REST API testing tools. This talk offered an in-depth look into the evolution of REST API testing and our proposed solution. You can find the relevant paper through this link and slides through this link.
Talk, The 2018 International Summer Undergraduate Research Fellowship at the University of California, Irvine, Irvine, California, United States
In this talk, I introduced Sayme, a chatbot that our team developed to detect and resolve potential code conflicts in parallel software development. Sayme is designed to operate both proactively, alerting developers to actions that could create conflicting changes, and reactively, responding to inquiries about the state of different developers’ work and potential overlaps. You can find the related paper through this link.
Poster, 2018 Korea Computer Congress, Je-Ju International Convention Center, Jeju Special Self-Governing Province, South Korea
In this presentation, I introduced a unique vulnerability detection system based on the Ethereum network. Blockchain, underpinned by asymmetric encryption techniques, enables sharing of distributed records on a peer-to-peer network. Its potential extends beyond cryptocurrency, with platforms like Ethereum, which can execute code, receiving considerable attention. However, ‘smart contracts’, where code is written in blockchains like Ethereum, can host vulnerabilities if the code contains errors. While tools such as Oyente, Securify, Mythril, MAIAN, and Gasper have been developed to mitigate these vulnerabilities, they rely on the original Solidity files for analysis, posing difficulties for non-developers or those analyzing others’ smart contracts. To address this, we implemented the Kookmin Security Token—a smart contract that serves as both an interface and data storage—and a connected analysis server. This novel system enables users to send the address of the smart contract they wish to analyze and receive results effortlessly, marking a significant step forward in the accessibility of blockchain security.
Talk, The 2017 International Summer Undergraduate Research Fellowship at the University of California, Irvine, Irvine, California, United States
In this talk, I introduced a novel approach to detecting social engineering attacks uniquely focuses on analyzing the natural language text within potential attacks, using semantic analysis to identify malicious intent. Unlike prior research that heavily relied on email metadata, our method is applicable across various communication platforms, including text and chat apps, as well as phone or in-person conversations converted to text. The effectiveness of our approach is demonstrated through extensive evaluation using a substantial benchmark set of phishing emails. You can find the relevant paper through this link.