Me

I am Xiaoyong (Brian) Yuan,

Fourth-year Ph.D student of Computer & Information Science & Engineering Department, University of Florida.

My advisor is Dr. Xiaolin (Andy) Li. I joined Andy's Scalable Software Systems Laboratory (S3Lab) in Fall, 2015. I also closely work with Dr. Daniela Seabra Oliveira. Before joining the Gator Nation, I received my M.E. in Computer Engineering from Peking University (2015), advised by Dr. Ying Li, and B.S. in Mathematics from Fudan University (2012).

My research interests include Deep Learning, Security, and Distributed Systems. CV

I come from Shanghai, a modern and beautiful city in China. When I'm not working, I like a lot of sports (e.g. Soccer, Swimming, Badminton, Basketball, Sanshou, Shengji, and Skiing). I'm a big fan of Shanghai Shenhua and A.C. Milan.

“Heaven has made us talents; we're not made in vain.” -BaiLi

天生我材必有用,千金散尽还复来。 李白

Research

  • Generalized Batch Normalization

    Utilizing recently introduced concepts from statistics and quantitative risk management, we present a general variant of Batch Normalization (BN) that offers accelerated convergence of Neural Network training compared to conventional BN. In general, we show that mean and standard deviation are not always the most appropriate choice for the centering and scaling procedure within the BN transformation, particularly if ReLU follows the normalization step. We present a Generalized Batch Normalization (GBN) transformation, which can utilize a variety of alternative deviation measures for scaling and statistics for centering, choices which naturally arise from the theory of generalized deviation measures and risk theory in general. When used in conjunction with the ReLU non-linearity, the underlying risk theory suggests natural, arguably optimal choices for the deviation measure and statistic. Utilizing the suggested deviation measure and statistic, we show experimentally that training is accelerated more so than with conventional BN, often with improved error rate as well. Overall, we propose a more flexible BN transformation supported by a complimentary theoretical framework that can potentially guide design choices.
  • DeepMalware

    Deep Models and Mechanisms for Malware Detection and Defense. A system call based malware detection algorithm + multi-stage online defense mechanisms.
  • DeepCloud

    We design and implement building blocks, services, and platforms for the deep learning research and big data applications with comprehensive expertise and experiences in deep learning, machine learning, systems, and applications. [Video]
  • DeepDetection: Identifying DDoS Attack via Deep Learning

    Distributed Denial of Service (DDoS) attacks grow rapidly and become one of the major threats to the Internet. Automatically detecting DDoS attack packets is one of the main defense mechanisms. Conventional solution monitors network traffic and identifies attack activities from legitimate network traffic based on statistical divergence. Machine learning is another method to improve identifying performance based on statistical features. However, conventional machine learning techniques are limited by the shallow representation models. We propose a deep learning based DDoS attack detection approach (called DeepDetection). Deep learning approach can automatically extract high level features from low level ones and gain powerful representation and inference. We build a recurrent deep neural network to learn patterns from sequences of network traffic and trace network attack activities in both long and short terms.

Publications

  • Xiaoyong Yuan*, Zheng Feng* (*equal contribution), Matthew Norton, Xiaolin Li, “Generalized Batch Normalization: Towards Accelerating Deep Neural Networks”, 2019, accepted by AAAI-19, (acceptance rate 16.2%).
  • Xiaoyong Yuan, Pan He, Xiaolin Li, “Adaptive Adversarial Attack on Scene Text Recognition”, 2018. arXiv
  • Xiaoyong Yuan, Pan He, Qile Zhu, Xiaolin Li, “Adversarial Examples: Attacks and Defenses for Deep Learning”, 2017, accepted by IEEE Transactions on Neural Networks and Learning Systems (TNNLS). arXiv, Github Resource List
  • Ruimin Sun*, Xiaoyong Yuan* (*equal contribution), Pan He, Qile Zhu, Aokun Chen, Andre Gregio, Daniela Oliveira, Xiaolin Li, “Learning Fast and Slow: PROPEDEUTICA for Real-time Malware Detection”, 2017. arXiv
  • Ruimin Sun, Xiaoyong Yuan, Andrew Lee, Matt Bishop, Donald E Porter, Xiaolin Li, Andre Gregio, Daniela Oliveira, “The Dose Makes the Poison--Leveraging Uncertainty for Effective Malware Detection”, IEEE Conference on Dependable and Secure Computing (DSC). Taipei, Taiwan, August 7-10, 2017.
  • Xiaoyong Yuan, Chuanhuang Li, Xiaolin Li, “DeepDefense: Identifying DDoS Attack via Deep Learning”, 3rd IEEE International Conference on Smart Computing (SMARTCOMP 2017), 29-31 May, Hong Kong, China, 2017.
  • Xiaoyong Yuan, Long Wang, Tiancheng Liu and Yue Zhang, “A Methodology for Continuous Evaluation of Cloud Resiliency”, American Journal of Engineering and Applied Sciences, 2016.
  • Xiaoyong Yuan, Hongyan Tang, Ying Li, Tong Jia, Tiancheng Liu, Zhonghai Wu “A Competitive Penalty Model for Availability based Cloud”, Services Transactions on Cloud Computing (STCC), 4(1), 2016, pp. 1-14.
  • Tang Hongyan, Li Ying, Jia Tong, Yuan Xiaoyong, Wu Zhonghai, “Time Series Based Killer Task Online Recognition Service: A Google Cluster Case Study”, the 10th International Conference on Service-Oriented System Engineering (IEEE SOSE 2016), Oxford, UK, Mar. 29 – Apr. 1, 2016. PDF
  • YUAN Xiaoyong, TANG Hongyan, LI Ying, JIA Tong, LIU Tiancheng, WU Zhonghai, “A Competitive Penalty Model for Availability Based Cloud SLA”, The 8th IEEE International Conference on Cloud Computing (Cloud 2015), New York, US, Jun. 27 - Jul. 2, 2015.
  • Jia Tong, LI Ying, Yuan Xiaoyong, Tang Hongyan, Wu Zhonghai, “Characterizing and Predicting Bug Assignment in OpenStack”, The 2nd International Conference on Trustworthy Systems and Their Applications (TSA 2015), Taiwan, Jul. 8-9, 2015.
  • YUAN Xiaoyong, LI Ying, JIA Tong, LIU Tiancheng, WU Zhonghai, “An Analysis on Availability Commitment and Penalty in Cloud SLA”, The 39th Annual International Computers, Software & Applications Conference (COMPSAC 2015), Taiwan, Jul. 1-5, 2015. PDF
  • YUAN Xiaoyong, LI Ying, WU Zhonghai, LIU Tiancheng, “Dependability Analysis on OpenStack IaaS Cloud: Bug Analysis and Fault Injection”, The 6th IEEE International Conference on Cloud Computing Technology and Science (CloudCom 2014), Singapore, Dec. 15-18, 2014. PDF
  • YUAN Xiaoyong, LI Ying, WANG Yanqi, SUN Kewei, “Scheduling Cloud Platform Managed Live-Migration Operations to Minimize the Makespan”, The 11th IFIP International Conference on Network and Parallel Computing (NPC 2014), Taiwan, Sep. 18-20, 2014.

Professional Services

  • Reviewer: Transactions on Parallel and Distributed Systems, IEEE Transactions on Evolutionary Computation, ICMLDS 2017, ICMLDS 2018, CCF Transactions on Networking
  • SubReviewer: S&P 2019, ACSAC 2017, ACSAC 2018, INFOCOM 2016, INFOCOM 2017
  • Experience

    Teaching Assistant

    • EEL6935 Big Data Ecosystems, Spring 2018: Pytorch Tutorial (slides)
    • CIS4301 Information and Database Systems I, Fall 2015

    Internship

    • IBM China Research Laboratory (Aug. 2013 - Jun. 2014)

    Awards

    • Best Poster Award of 2018 FICS Research Conference on Cybersecurity
    • NSF student travel grant of 3rd IEEE Conference on Smart Computing (SMARTCOMP 2017)
    • Student Travel Award of IEEE ICWS/SCC/CLOUD/MS/BigDataCongress/SERVICES 2015

    Services

    • Lead Organizer of UF Artificial Intelligence and Security Club (AISec)
    • Volunteer teaching in Shandong Province as a vice group leader (Jul. 2010)

    Contact

    Email

    chbrian-at-ufl-dot-edu

    Address

    1064 Center Dr. NEB 406, Gainesville, Florida 32611

    Homepage

    http://www.cise.ufl.edu/~xiaoyong/