D , high-performance software has always been difficult. Two contemporary challenges exacerbate matters further: the ubiquity of multi-core processors, which demand parallel software, and the increasing size and intricacy of software projects. My research program is motivated by my first-hand experience of these challenges as a developer and researcher. I seek to () raise the a...

Even though advances in hardware technologies are constantly pushing the limits of computational and storage resources, there are always applications whose computational demands exceed even the fastest technologies available. Hence, it has become critical to look into ways to aggregate efficiently distributed resources to benefit a single application. Achieving this vision requires being able t...

My dissertation research developed reliable and available servers capable of running unmodified shared-memory applications. In collaboration with my colleagues at CMU, I proposed the TRUSS (Total Reliability Using Scalable Servers) architecture [3], which allows designers to separately address processor, memory, and I/O reliability. My thesis work shows that redundant processor pairs can be dis...

My general research interests are within the area of text information management. Text information, notably the various types of online information, is continuously produced everywhere in every possible way. The dramatic growth of text information has brought changes to many aspects of one’s daily life. Even in our own research, online text information, such as scientific literature and technic...

A multidisciplinary research background I earned my Ph.D. in Information Science at Rutgers University, M.A. degree in Computermediated Communication at University of Florida, and B.A. in Management Information Systems at National Sun-Yat-Sen University in Taiwan. My multidisciplinary education backgrounds provide me a unique and novel perspective in studying information systems and enable me t...

My general research interests lie in the statistical inference of stochastic processes superimposed on stochastically evolving networks. Stochastic processes are broadly defined to include Poisson processes, Markov chains, and non-linear stochastic differential equations. Stochastically evolving networks are meant to provide the biologically realistic contexts within which our stochastic proces...

When we integrate data from multiple sources, the resulting database may contain errors and inconsistencies. Ideally, all errors should be cleaned automatically. In reality, though, data cleaning necessarily requires human intervention. To facilitate this, I envision a data integration process where it is possible to obtain meaningful and consistent answers from an integrated database even if i...

Figure 1: Protein 3W6H in wireframe (a) and in spheres (b) viewed by PyMOL. The protein is modeled as a rigid obstacle (c) and we can use a recently developed method which guarantees the sample distribution around the obstacle to generate possible binding ligand configurations with respect to the target protein (d). My graduate research focus is motion planning [4, 6] and its application in com...