Chi Zhang

Understanding User’s Preferences for Service Selection: During our exploration of the key leakage project, we found many developers face the problem of how to select a proper service for their applications. To better understand developers’ choices and help service developers enhance their services, we conduct an empirical study on how different characteristics of services impact developers' choices. We studied the commit logs in GitHub to identify API replacements, and manually labeled what APIs can be used to subtitle the APIs already used in repositories. We plan to submit the work to ICWS’24.

API Marketplace Key Leakage Analysis: In this project, we identified security issues with API marketplaces such as RapidAPI, measured its impact, and proposed potential solutions. RapidAPI is a multi-cloud, customized hub to find, connect, and share APIs, with more than 4 million developers and 40K available APIs. By crawling and analyzing Github repositories that invoke APIs from the platform, we found that more than 40,000 keys to access various APIs are exposed as plaintext, which can be used to perform DDoS, Theft of Service attacks. We further studied the impact of such key leakages, and proposed several solutions to detect and prevent such attacks. We plan to submit the work to AsiaCCS’24.

Discrimination in web services: For my undergraduate thesis, I worked on identifying the discrimination issues in web services. We found that many services give inaccurate results for certain user inputs. For example, some facial recognition services may experience a significant lower accuracy for facial images of asian and females. To identify such services, we developed a sampling mechanism, which chooses representative samples from a whole set, to verify the output of services. Our goal is to minimize the number of samples while achieving considerable accuracy.

Shandong University-Nanyang Technological University Artificial Intelligence International Joint Research Institute project, participated in the construction of a knowledge map of Chinese medicine direction.

• "EasyTrans Translation"-Automatic Audio and Video Summary app Based on Deep Learning
• Responsible for demand positioning analysis, control the overall project process; Mainly responsible for back-end development. This project provides an automatic generation scheme of abstracts and notes for knowledge-intensive online class videos. It won the "Challenge Cup" special prize for college students' extracurricular academic and technological works, the first prize of Shandong University Student Software Design Competition, and applied for 2 patents (Accepted)