This work coincides with Microsoft Corporation marking the tenth anniversary of Microsoft Research Asia (MSRA), its largest research and development facility outside of the US. The anniversary recognizes through Innovation Day in Beijing that the past ten years have been characterized by the creativity of Chinese research and the enthusiasm of Microsoft to leverage the expertise found in the top universities of China and the rest of Asia. The MSRA facility has five areas of concentration: data-intensive computing, next-generation multimedia, computer science fundamentals, search and online advertising, and natural user interfaces. New areas of research are also in the offing, in fields such as cloud computing and software as a service (SaaS).
The volume deals with most of these areas except, of course, the emerging fields. The research in the book is indicative of the increasing innovative nature of Chinese computing, as China is transformed from a manufacturing powerhouse to an innovation contender.
The contents of the book include four main areas: next-generation multimedia, networking and systems, natural user interfaces, and search and mining. Within each of the four areas, there are between two to seven essays.
The degree of innovation and interest in next-generation multimedia may be inferred from the length of this particular section, since there are seven essays in this area. In fact, just over half of the book addresses issues of next-generation multimedia. On the other hand, networking elicited four essays, while user interfaces and search and mining received treatment in two essays each. The emphasis, then, is on next-generation multimedia.
A sampling of the next-generation multimedia section illustrates the innovative nature of these contributions. The several authors of “Drag and Drop Pasting” propose “a user-friendly approach to achieve seamless image composition without requiring careful initialization by the user.” The authors of a second essay propose “motion patches,” as building blocks for virtual environments and computer games. The authors of “Video Completion by Motion Field Transfer” suggest filling in “missing video parts by sampling spatio-temporal patches of local motion instead of directly sampling color,” as is currently typical. The researchers of “Correlative Multi-Label Video Annotation” propose “a correlative multi-labeling (CML) approach to exploit how the concept correlations help infer the video semantic concepts.” The writers of “Stereoscopic Video Synthesis from a Monocular Video” offer that a “convincing and smooth stereoscopic video can be synthesized even by simple constant-depth warping,” although a user may only be using an ordinary handheld video camera. “Shape Palettes: Interactive Normal Transfer via Sketching” outlines “a simple interactive approach to specify 3D shape in a single view using ‘shape palettes.’” Finally, “Adaptive Directional Lifting-based Wavelet Transform for Image Coding” suggests that adaptive directional lifting (ADL) “performs lifting-based prediction in local windows in the direction of high pixel correlation.”
Networking is the next most important area of interest, and two of the papers are particularly intriguing. The first, “Low-Power Distributed Event Detection in Wireless Sensor Networks,” maintains that a completely localized algorithm “significantly improves detection performance.” “Understanding User Behavior in Large-Scale Video-on-Demand Systems” reveals the surprising result “that video session length has a weak inverse correlation with the video’s popularity.” A final essay in this section addresses an important area of cryptographic protocols. Part 3 addresses the important user interface issues of voice and speech recognition, and Part 4 illuminates Web query classification and spectral domain-transfer learning.
It is clear that Microsoft Asia has numerous fulfilling research paths to follow; I would expect further quality research to emerge from the center. Ongoing research is needed, and future work at MSRA may be in the burgeoning fields. The increased demand for efficient data centers and computers is but one example: more than 1.5 percent of all energy is being used by computers, and sensors developed by Microsoft have helped reduce cooling and energy consumption by over 30 percent where they were implemented.
