基本情况：郭同学，University of Hong Kong，B.Eng. Computer Engineering，GPA 3.76，TOEFL 105，GRE 325
录取学校：University of Illinois at Urbana Champaign Ph.D. in Computer Science (全奖 $61600 美金)
The rapid development of computing power increased the attention afforded to computer science. This challenging field will bring dramatic changes to lifestyles despite possible difficulties. As a person who takes delight in solving difficulties, I hope to explore problems in networks and build robust and reliable network systems.
Grew up in an academic family, I am also ready to devote myself in teaching at university, conduct research focus on network and systems, and even set up my own laboratory. To obtain doctoral degree and accumulate valuable research experiences at UIUC are critical to the realization of my goal. My one-year exchange experience at UIUC exposed me to the outstanding progress the Computer Science Department have accomplished. This experience greatly inspired me and motivated me to pursue doctoral research at UIUC, and conduct researches as follows:
Mobile and Ubiquitous Computing and Wireless Sensor Network. My research interest includes mobile or wireless applications on networking and sensing. Numerous mobile devices are connected to networks and the number of mobile users will continue to grow. My research vision is to make fast and robust wireless mobile networks to make everyday things smart and connected. Besides, I also wish to develop emerging ubiquitous applications, such as mobile sensing and wireless localization.
Network Economics and Algorithmic Game Theory. I am interested in applying game theory to create social products in network systems. In particular, I am currently working on and will further focus on incentive mechanisms designing for mobile crowd sensing systems, which would serve a wide spectrum of applications in healthcare and environmental monitoring. I am also interested in designing online algorithms in distributed systems to achieve optimal object value, such as resource allocation and management of cloud computing systems, and multi-queuing scheduling of wireless networks.
Datacenter Networks and Cloud Computing. I am interested in building fast and robust network for data centers and cloud computing systems. The huge amount of data flowing in and out of data centers resulted in routing, traffic management, and load balancing maintenance problems. The development of software-defined networking facilitates the possible integration of a central control in a distributed system and sheds light on obstacles of data center networks. Excitement fills me at the thought of conducting relevant research and making contribution to SDN.
Reinforcement Learning for Networks and Systems. My vision of applying machine learning methods to networks and systems is based on the availability of big data in networks and emerging data analysis/machine learning methods. My goal is to develop reinforcement learning techniques to solve problems in networks and systems, such as virtual machine scheduling in data center and real-time bidding for on-line advertisement auction. Finding the exact representation of such problems and deriving optimal value functions are challenging. These problems can be addressed by approximating value functions on a smaller state space with reinforcement learning and deriving tight bounds for such approximation. Exploring and learning from huge data could assist in designing reliable and efficient network systems.
UIUC is the holy place for my dreams. The research activities in the Systems and Networking group are in line with mine. Particularly, I am interested in Professor Klara Nahrstedt, Professor Tarek Abdelzaher, and Professor Haitham Hassanieh’s research work. Finishing my Ph.D. in Computer Science program study at UIUC will undoubtedly actualize my academic and career goals.
I have obtained competitive academic preparation, and also seized the invaluable opportunities to develop my research interest. On account of my academic performance, I gained entry into the Undergraduate Research Fellowship Program (top 5%), finished my research at UIUC and published my papers. At UIUC, I joined Prof. Klara Nahrstedt’s work on Multimedia Operating Systems and Networking Group (MONET). Ourresearch centered on building a stable wireless sensor network with high efficiency and high rate of fault tolerance to realize data collection and environment monitoring.
During the research, I conducted largely independent work on the design and optimization of the resilient data collection protocol, which would preprocess collected data from the sensor, and thus reduce unnecessary data transmission and find problems in the environment. The research comprised three objectives. First, building an efficient ad hoc data reporting structure in the network to quicken data collection. Second, organizing and preprocessing the network data to achieve on-siting environment monitoring. Third, enabling the system to identify alternative relies for disconnected children sensors and thus avoid massive data loss when some sensors failed in the network.
Separating each component was easy, but ensuring that each part functions simultaneously was difficult. Fortunately, reading substantial literature and referring to related research enabled me to introduce a tree-based multi-sink structure to improve the speed of data collection. I incorporated quorum scheme to gather highly-correlated data in network for preprocessing based on the geographical features of the sensors. Finally, the resiliency issue was addressed by exploiting optimal redundant connections in the sensor networks.
The results of this project have been published as a workshop paper wherein I am the first author. Through this independent research experience, I fostered my ability to propose questions, search for solutions, and collaborate with others, as well as express ideas and write papers. Also, in the process, I have realized that the essence of solving problems in network lies in exploring ways to distribute limited resources under various restrictions to obtain the maximum effect, and also the distributed feature of network augments the difficulty of such kind of optimization problems. I am interested in tackling those challenging tasks. My interest in building faster and more stable network system is also intensified in this project.
My experiences demonstrate my capability of taking new challenges and finishing doctoral study at UIUC. Through my experience in research and as laboratory assistant, I have improved my communication skills with teachers and peers. If admitted, I am confident that I will quickly adapt to the research work and contribute to the department. Thus, admission to UIUC will set me on track in career.