- Jun Li
Despite intensive research over 10 years, Android malware detection still faces substantial challenges including frequent changes in the Android framework, the existence of noisy labels in large-scale up-to-date datasets, and the continuous evolution of Android malware. The consequences of ignoring these challenges are multifold. One is the fast decline of malware detection accuracy over time due to the use of out-of-date malware detection feature sets, the ignorance of new APIs and changed APIs, and the failure of capturing emerging malware patterns. On the other hand, the use of noisy datasets leads to distorted training of malware detection models, unfair evaluation of malware detection performance, and unidentified false positives and false negatives. To address these challenges, this talk summarizes three robust malware detection approaches, including DroidEvolver, SDAC, and Differential Training. Rigorous experiments show that these approaches improve the robustness of the state-of-art approaches significantly.
Yingjiu Li is currently a Ripple Professor in the Department of Computer and Information Science at the University of Oregon. His research interests include IoT Security and Privacy, Mobile and System Security, Applied Cryptography and Cloud Security, and Data Application Security and Privacy. He has published over 150 technical papers in international conferences and journals, and served in the program committees for over 80 international conferences and workshops, including top-tier cybersecurity conferences.