Ph.D. Candidate Department of Electrical Engineering and Computer Science Syracuse University Office: Science&Technology Center 4-226, Syracuse University Address: 50 Presidential Plz 214, Syracuse, NY, 13202 Email: qchen14 (at) syr (dot) edu

Profile, Education, Research Projects, Internship, Publications, Other Projects

PROFILE

I am currently a research assistant at the AMPS Lab at Syracuse University. My advisor is Dr. Qinru Qiu. Generally, my research interests are machine learning, cognitive applications and high performance computing on heterogeneous systems. In particular, I have been working on real-time anomaly detection using inference model and their GPU accelerations. I am currently working on visual-semantic modeling between deep image features and loosely correlated texts.

EDUCATION

• Ph.D. in Computer Engineering, Syracuse University, Present
• M.S. in Electrical Engineering, Beijing University of Posts & Telecoms (BUPT), 2012
• B.S. in Electrical Engineering, BUPT & Queen Mary University of London, 2009
  First class honours degree

RESEARCH PROJECTS

Bidirectional mapping between deep image and noisy texts Design visual-semantic retrieval model for loosely coupled text-image pairs such as picture news. Cascade fragment embedding stages to remove noisy texts and improve learning efficiency. Integrate deep face network for better news event correlation.

Neuromorphic framework for real-time anomalous stream detection
Develop a self-structuring knowledge network that provides continuous traffic monitoring and prompt anomaly detection/reasoning for massive concurrent data streams. Implement on heterogeneous multicore platform with GPGPUs and Xeon Phi co-processors.

Context-aware smartphone low power design
Learn battery consumption patterns from user behaviors to support control decisions that save device power while maintaining QoS.

Intelligent Text Recognition System
Study statistical inference based sentence completion model, coupled with BSB pattern matching to provide machines strong reading/synthesizing ability.

INTERNSHIPS

Microsoft, Redmond, WA, USA, May 2015 - Aug 2015 SWE Intern: Develop intelligent retransmission pipeline for Skype media stack. Participate Bing Ads click-through prediction hachathon.

Alcatel-Lucent, Beijing, China, Jan 2012 - Apr 2012
SWE Intern: Develop and maintain wireless network elements management software. Ported C/C++ programs from Solaris to Linux servers.

Potevio Institute of Technology, Beijing, China, May 2010 - Jan 2011
Baseband Engineer Intern: Develop a TD-LTE downlink demo platform and conducted system integration tests. Implement baseband processing modules and an on-chip protocol stack.

PUBLICATIONS

1. Qiuwen Chen and Qinru Qiu, “Enhancing Bidirectional Association between Deep Image Representations and Loosely Correlated Texts”, IJCNN’2016. pdf
2. Qiuwen Chen, Qing Wu, Morgan Bishop, Richard Linderman and Qinru Qiu, “Self-structured Confabulation Network for Fast Anomaly Detection and Reasoning”, IJCNN’2015. pdf
3. Qiuwen Chen, Qinru Qiu, Qing Wu, Morgan Bishop, and Mark Barnell, “A Confabulation Model for Abnormal Vehicle Events Detection in Wide-Area Traffic Monitoring”, CogSIMA’2014. pdf
4. Q. Wu, B. Liu, Y. Chen, H. Li, Qiuwen Chen and Qinru Qiu, “Bio-Inspired Computing with Resistive Memories - Models, Architectures and Applications”, ISCAS’2014. pdf
5. Hao Shen, Qiuwen Chen and Qinru Qiu, “Battery Aware Stochastic QoS Boosting in Mobile Computing Devices”, DATE’2014. pdf
6. Qiuwen Chen, Qinru Qiu, Hai Li and Qing Wu, “A Neuromorphic Architecture for Anomaly Detection in Autonomous Large-Area Traffic Monitoring”, ICCAD’2013. pdf
7. Qiuwen Chen, Z. Wu, Hao Jin, “A Multi-user Cooperation Scheme based on Component Demodulation”, ICCTA’2011. pdf
8. Qiuwen Chen, “Enhanced Turbo Network Coding with Additional Error Protection by Hybrid Iterative Decoding”, ICCSN’2011. pdf

OTHER PROJECTS

Friend recommendation service in distributed social network: implemented a basic friend recommendation module on Diaspora. Further study the policy extension using simulation on Facebook dataset. pdf src

Cross-proximity probabilistic modeling of actor behavior: designed a link prediction algorithm for dynamic social networks based on a dual hidden Markov model (HMM) coupled with diversified influence factors. pdf src

Malware detection from system call sequences: developed and evaluated learning based models such as replicator neural network, CFA and Bayesian net on identifying malicious programs.

GPU acceleration of BCJR algorithm: originated an augmented parallel algorithm that exploits the thread level parallelism. Verified the performance gain on GPUs. pdf

Resource usage prediction for cloud data center: built a confabulation based statistic prediction model for virtual machine resource consumptions (clock cycles, memory, networks). pdf

VLSI computer aided design: Implemented software with GUIs for a set of CAD algorithms: layout optimization, GORDIAN placement, simulated annealing, etc. src

Joint coding modulation diversity: Focused on adapting hybrid diversity techniques to aid wireless transmissions in fading channels. Designed a turbo receiver which harvests signal space diversity and coding gain jointly. Implemented a testbed on PicoChip many-core system.

Multi-user signal space cooperation: Developed a wireless cooperation scheme in uplink transmissions. Designed a 4D component demodulation method that enables multiple users exploiting path diversity in a cross-layered manner. src

Microcontroller-based Contour Mapper: hardware & software co-design of a conceptual sensor robot that collects coordinate information and plot the contour map.