Professor, student win best paper award at prestigious international data science forum

 Yanfang "Fanny" Ye and graduate student Shifu Hou captured top honors at the Conference on Knowledge Discovery and Data Mining.

Representatives from West Virginia University’s Statler College of Engineering and Mineral Resources captured best paper honors at the premier international forum for data science.

Assistant Professor of Computer Science and Electrical Engineering Yanfang “Fanny” Ye and doctoral student Shifu Hou received the prestigious KDD 2017 Best Paper and Best Student Paper awards in the applied science track at the annual Conference on Knowledge Discovery and Data Mining, held recently in Halifax, Nova Scotia. The conference, which brings together researchers and practitioners from data science, data mining, knowledge discovery, large-scale data analytics and big data, only accepts 10 percent of the papers submitted for presentation.

The paper, “HinDroid: An Intelligent Android Malware Detection System Based on Structured Heterogeneous Information Network,” touted the team’s novel feature presentation based on heterogeneous information network—or HIN—which can push malware detection on Android devices to almost 100 percent.

According to the paper HinDroid had a 98.6 per cent recognition rate in the lab. By comparison, in tests with identical malware samples other techniques had recognition rates varying from 88.6-95.2 percent. The research is being funded in part by the National Science Foundation and New Jersey-based security vendor Comodo Group, which is testing the technique for possible use in its cloud-based enterprise mobile security service.

“With the explosive growth of malware and the severity of its damages to smart phone users, the detection of Android malware has become increasingly important in cybersecurity. The increasing sophistication of Android malware calls for new defensive techniques that are capable against novel threats and harder to evade,” said Ye. “Instead of only using Application Programming Interface calls, we further analyzed the different relationships between them and created higher-level semantics that require more efforts for attackers to evade the detection.

“We represented the Android applications, related APIs and their rich relationships as a structured heterogeneous information network,” Ye continued. “Then we used a meta-path-based approach to characterize the semantic relatedness of apps and APIs, and the aggregate different similarities using multi-kernel learning by which each meta-path is automatically weighted to make predictions. To the best of our knowledge, this is the first work to use structured HIN for Android malware detection.”

Ye has extensive research and development experience in Internet security solutions. She proposed and developed cloud-based solutions for mining big data in the area of Internet security, especially for malware detection and phishing fraud detection. Her developed algorithms and systems have been incorporated into popular commercial products, including Comodo Internet Security with more than 35 million users in the U.S., and Kingsoft Antivirus, with 150 million-plus users in China.

Yangqiu Song, assistant professor of computer science and engineering at Hong Kong University of Science and Technology and a former assistant professor at WVU, collaborated on the paper.

---