Page 173 - My FlipBook
P. 173
Research Fellow 陳孟彰 研
究
Meng Chang Chen 人
員
Ph.D., Computer Science, University of California at Los Angeles, United States Faculty
T +886-2-2788-3799 ext. 1802 E mcc@iis.sinica.edu.tw
F +886-2-2782-4814 W www.iis.sinica.edu.tw/pages/mcc
・ Deputy Director, Institute of Information Science, Academia Sinica (1999-2002, 2008-
2010)
・ Principal investigator, Patent and Technology Exchange Group, National Telecom
Project O ce, (2001-2003)
・ Member of Technical Sta , Bell Labs, AT&T, (1989-1992)
Research Description
My research interests are in computer and network software system, and data and knowledge engineering, and various problems within
the areas. In the 2000s, I worked on text mining problem, especially in summarization and story generation from a collection of articles (e.g.,
query return from Google). Simultaneously, I also worked on the wireless networking, including multi-hop mesh transmission, 5G small cell
ultra-dense network, and network mobility on high speed vehicle.
Currently, I work on two issues, deep learning for PM2.5 prediction, and malware analysis. First, we proposed the composite neural network
framework, which proves with high probability that a composite neural network performs better than any of its pre-trained component,
regardless of topology. The PM2.5 prediction is a great challenge that requires both deep domain knowledge and deep learning skills. We,
based on composite neural network framework, considered many highly in uencing factors (e.g., remote transportation, time-e ect) and
developed needed theories (e.g., extreme case distribution, step-function-like loss function) in order to obtain high quality ne-grained 72
hours prediction.
For malware analysis, rst, we built a hardware-assisted VMI (virtual machine introspection) mechanism from a software stack containing
virtual machine, hypervisor, and QMEU and run captured malware in this control environment. Our malware analysis includes both static
and dynamic analyses. The dynamic analysis uses the execution trace collected from the rst step to further investigate the API/system calls
and resources used by the malware. The static analysis looks into the malware program to perform statistics on API calls and generates call
graph and extract the low frequency calls. Later on, we considered all the extractions as features and applied the composite neural network
framework to build an e cient and high accuracy deep learning apparatus for various malware analysis, such as malware family classi cation.
Currently, we use Mitre Att&ck repository to extract Malware attack knowledge to identify the attack path of a malware.
1. "Composite Neural Network: Theory and Application to Publications Brochure 2020
PM2.5 Prediction", Ming-Chuan Yang and Meng Chang Chen,
arXiv:1910.09739, Oct., 2019. 6. "A Two-Stage Link Scheduling Scheme for Variable-Bit-Rate
Traffic Flows in Wireless Mesh Networks", Yung-Cheng Tu,
2. "Hardware-Assisted MMU Redirection for In-guest Monitoring Meng Chang Chen and Yeali S. Sun, IEEE Transactions on
and API Profiling", Mike Hsiao, Yeali Sun, Meng Chang Chen, Wireless Communications, November 2014.
accepted by IEEE Transactions on Information Forensics &
Security. 7. "Coding-Aware Peer-to-Peer Data Repair in Multi-Rate Wireless
Networks - A Game Theoretic Analysis", Hsiao-Chen Lu,
3. "NASH: Navigation-Assisted Seamless Handover Scheme for Wanjiun Liao, Meng Chang Chen, and Musaed A. Alhussein,
Smart Car in Ultra-Dense Networks", Ming-Chin Chuang, Meng IEEE Journal on Selected Areas in Communications, Special
Chang Chen, IEEE Transactions on Vehicular Technology , Issue on Emerging Technologies in Communications, September
February 2018. 2013.
4. "Slow-Paced Persistent Network Attacks Analysis and Detection 8. "Protocol and Architecture Supports for Network Mobility with
Using Spectrum Analysis", Li Ming Chen, Shun-Wen Hsiao, QoS-Handover for High-Velocity Vehicles", Cheng-Wei Lee,
Meng Chang Chen, and Wanjiun Liao, IEEE Systems Journal, Yeali Sun and Meng Chang Chen, Wireless Networks. July 2013.
December 2016.
9. "TSCAN: A Content Anatomy Approach to Temporal Topic
5. "Seamless Handover for High-Speed Trains Using Femtocell- Summarization", Chien Chin Chen and Meng Chang Chen, IEEE
based Multiple Egress Network Interfaces", Cheng-Wei Lee, Transactions on Knowledge and Data Engineering, January 2012.
Ming-Chin Chuang, Meng Chang Chen and Yeali S. Sun, IEEE
Transactions on Wireless Communications, December 2014. 10. "Using Incremental PLSI for Threshold Resilient Online Event
Analysis", Tzu-chuan Chou and Meng Chang Chen, IEEE
Transactions on Knowledge and Data Engineering, March 2008.
171
究
Meng Chang Chen 人
員
Ph.D., Computer Science, University of California at Los Angeles, United States Faculty
T +886-2-2788-3799 ext. 1802 E mcc@iis.sinica.edu.tw
F +886-2-2782-4814 W www.iis.sinica.edu.tw/pages/mcc
・ Deputy Director, Institute of Information Science, Academia Sinica (1999-2002, 2008-
2010)
・ Principal investigator, Patent and Technology Exchange Group, National Telecom
Project O ce, (2001-2003)
・ Member of Technical Sta , Bell Labs, AT&T, (1989-1992)
Research Description
My research interests are in computer and network software system, and data and knowledge engineering, and various problems within
the areas. In the 2000s, I worked on text mining problem, especially in summarization and story generation from a collection of articles (e.g.,
query return from Google). Simultaneously, I also worked on the wireless networking, including multi-hop mesh transmission, 5G small cell
ultra-dense network, and network mobility on high speed vehicle.
Currently, I work on two issues, deep learning for PM2.5 prediction, and malware analysis. First, we proposed the composite neural network
framework, which proves with high probability that a composite neural network performs better than any of its pre-trained component,
regardless of topology. The PM2.5 prediction is a great challenge that requires both deep domain knowledge and deep learning skills. We,
based on composite neural network framework, considered many highly in uencing factors (e.g., remote transportation, time-e ect) and
developed needed theories (e.g., extreme case distribution, step-function-like loss function) in order to obtain high quality ne-grained 72
hours prediction.
For malware analysis, rst, we built a hardware-assisted VMI (virtual machine introspection) mechanism from a software stack containing
virtual machine, hypervisor, and QMEU and run captured malware in this control environment. Our malware analysis includes both static
and dynamic analyses. The dynamic analysis uses the execution trace collected from the rst step to further investigate the API/system calls
and resources used by the malware. The static analysis looks into the malware program to perform statistics on API calls and generates call
graph and extract the low frequency calls. Later on, we considered all the extractions as features and applied the composite neural network
framework to build an e cient and high accuracy deep learning apparatus for various malware analysis, such as malware family classi cation.
Currently, we use Mitre Att&ck repository to extract Malware attack knowledge to identify the attack path of a malware.
1. "Composite Neural Network: Theory and Application to Publications Brochure 2020
PM2.5 Prediction", Ming-Chuan Yang and Meng Chang Chen,
arXiv:1910.09739, Oct., 2019. 6. "A Two-Stage Link Scheduling Scheme for Variable-Bit-Rate
Traffic Flows in Wireless Mesh Networks", Yung-Cheng Tu,
2. "Hardware-Assisted MMU Redirection for In-guest Monitoring Meng Chang Chen and Yeali S. Sun, IEEE Transactions on
and API Profiling", Mike Hsiao, Yeali Sun, Meng Chang Chen, Wireless Communications, November 2014.
accepted by IEEE Transactions on Information Forensics &
Security. 7. "Coding-Aware Peer-to-Peer Data Repair in Multi-Rate Wireless
Networks - A Game Theoretic Analysis", Hsiao-Chen Lu,
3. "NASH: Navigation-Assisted Seamless Handover Scheme for Wanjiun Liao, Meng Chang Chen, and Musaed A. Alhussein,
Smart Car in Ultra-Dense Networks", Ming-Chin Chuang, Meng IEEE Journal on Selected Areas in Communications, Special
Chang Chen, IEEE Transactions on Vehicular Technology , Issue on Emerging Technologies in Communications, September
February 2018. 2013.
4. "Slow-Paced Persistent Network Attacks Analysis and Detection 8. "Protocol and Architecture Supports for Network Mobility with
Using Spectrum Analysis", Li Ming Chen, Shun-Wen Hsiao, QoS-Handover for High-Velocity Vehicles", Cheng-Wei Lee,
Meng Chang Chen, and Wanjiun Liao, IEEE Systems Journal, Yeali Sun and Meng Chang Chen, Wireless Networks. July 2013.
December 2016.
9. "TSCAN: A Content Anatomy Approach to Temporal Topic
5. "Seamless Handover for High-Speed Trains Using Femtocell- Summarization", Chien Chin Chen and Meng Chang Chen, IEEE
based Multiple Egress Network Interfaces", Cheng-Wei Lee, Transactions on Knowledge and Data Engineering, January 2012.
Ming-Chin Chuang, Meng Chang Chen and Yeali S. Sun, IEEE
Transactions on Wireless Communications, December 2014. 10. "Using Incremental PLSI for Threshold Resilient Online Event
Analysis", Tzu-chuan Chou and Meng Chang Chen, IEEE
Transactions on Knowledge and Data Engineering, March 2008.
171
