Tue 03.12.19
11:00A EDT/4:43P PDT | 1 Hour Event
Tue 03.12.19
11:00A EDT/4:43P PDT
1 Hour Event
Tue 03.12.19
11:00A EDT/4:43P PDT | 1 Hour Event
Tue 03.12.19
11:00A EDT/4:43P PDT
1 Hour Event
Speaker
Ruimin Sun, PhD Candidate, Florida Institute of Cybersecurity, University of Florida
Location
655 Interdisciplinary Science & Engineering Complex (ISEC) 655
Abstract
This talk will introduce a promising avenue for improving the effectiveness of behavioral-based malware detectors is to combine fast (not highly accurate) machine learning methods with high-accuracy but time-consuming deep learning models. The main idea is to place software that receives borderline classifications from machine learning methods in an uncertain environment, and leverage deep learning models to do further analysis. The uncertainties will disproportionately affect poorly-written malware compared to well-written and resilient benign software, with the increasing demand for defensive programming. This talk will also discuss a statistical analysis of application sensitivity to OS uncertainty, which can help developers discover challenging bugs and identify the scenarios that most need validation. The results provided insights for software developers in building more resilient software.
Abstract
Ruimin is a PhD candidate in the Florida Institute of Cybersecurity at the University of Florida, advised by Professor Daniela Oliveira. Ruimin earned her bachelor’s degree at Southeast University in Nanjing, China before joining UF, where she studies robotics and measurement control technology. Ruimin’s research focuses on malware detection and software reliability in ubiquitous systems, and machine learning based performance analysis. Her future work will include information flow tracking and forensic analysis in IoT systems.
Speaker
Ruimin Sun, PhD Candidate, Florida Institute of Cybersecurity, University of Florida
Location
655 Interdisciplinary Science & Engineering Complex (ISEC) 655
Abstract
This talk will introduce a promising avenue for improving the effectiveness of behavioral-based malware detectors is to combine fast (not highly accurate) machine learning methods with high-accuracy but time-consuming deep learning models. The main idea is to place software that receives borderline classifications from machine learning methods in an uncertain environment, and leverage deep learning models to do further analysis. The uncertainties will disproportionately affect poorly-written malware compared to well-written and resilient benign software, with the increasing demand for defensive programming. This talk will also discuss a statistical analysis of application sensitivity to OS uncertainty, which can help developers discover challenging bugs and identify the scenarios that most need validation. The results provided insights for software developers in building more resilient software.
Abstract
Ruimin is a PhD candidate in the Florida Institute of Cybersecurity at the University of Florida, advised by Professor Daniela Oliveira. Ruimin earned her bachelor’s degree at Southeast University in Nanjing, China before joining UF, where she studies robotics and measurement control technology. Ruimin’s research focuses on malware detection and software reliability in ubiquitous systems, and machine learning based performance analysis. Her future work will include information flow tracking and forensic analysis in IoT systems.
Speaker
Ruimin Sun, PhD Candidate, Florida Institute of Cybersecurity, University of Florida
Location
655 Interdisciplinary Science & Engineering Complex (ISEC) 655
Abstract
This talk will introduce a promising avenue for improving the effectiveness of behavioral-based malware detectors is to combine fast (not highly accurate) machine learning methods with high-accuracy but time-consuming deep learning models. The main idea is to place software that receives borderline classifications from machine learning methods in an uncertain environment, and leverage deep learning models to do further analysis. The uncertainties will disproportionately affect poorly-written malware compared to well-written and resilient benign software, with the increasing demand for defensive programming. This talk will also discuss a statistical analysis of application sensitivity to OS uncertainty, which can help developers discover challenging bugs and identify the scenarios that most need validation. The results provided insights for software developers in building more resilient software.
Abstract
Ruimin is a PhD candidate in the Florida Institute of Cybersecurity at the University of Florida, advised by Professor Daniela Oliveira. Ruimin earned her bachelor’s degree at Southeast University in Nanjing, China before joining UF, where she studies robotics and measurement control technology. Ruimin’s research focuses on malware detection and software reliability in ubiquitous systems, and machine learning based performance analysis. Her future work will include information flow tracking and forensic analysis in IoT systems.
Speaker
Ruimin Sun, PhD Candidate, Florida Institute of Cybersecurity, University of Florida
Location
655 Interdisciplinary Science & Engineering Complex (ISEC) 655
Abstract
This talk will introduce a promising avenue for improving the effectiveness of behavioral-based malware detectors is to combine fast (not highly accurate) machine learning methods with high-accuracy but time-consuming deep learning models. The main idea is to place software that receives borderline classifications from machine learning methods in an uncertain environment, and leverage deep learning models to do further analysis. The uncertainties will disproportionately affect poorly-written malware compared to well-written and resilient benign software, with the increasing demand for defensive programming. This talk will also discuss a statistical analysis of application sensitivity to OS uncertainty, which can help developers discover challenging bugs and identify the scenarios that most need validation. The results provided insights for software developers in building more resilient software.
Abstract
Ruimin is a PhD candidate in the Florida Institute of Cybersecurity at the University of Florida, advised by Professor Daniela Oliveira. Ruimin earned her bachelor’s degree at Southeast University in Nanjing, China before joining UF, where she studies robotics and measurement control technology. Ruimin’s research focuses on malware detection and software reliability in ubiquitous systems, and machine learning based performance analysis. Her future work will include information flow tracking and forensic analysis in IoT systems.
