Wireless Computing in Medicine
Wireless Computing in Medicine
I recently celebrated my 50th birthday 26 productive years after I received my Ph.D. On this important milestone, I reflected back on my life, as I could not help but find myself in total agreement with what both Aristotle and Einstein said: the more one learns or knows, the more one realizes how much he/she does not know. I always wanted to learn more, so over the years I have expanded my parallel processing expertise from heterogeneous computing (topic of my first book) to bio-inspired and nanoscale integrated computing (topic of my second book). Expanding the application of both of these technologies further to medicine with an emphasis on their legal and ethical aspects is the main aim of this third book. This book is a product of the progression of my research, from undergraduate study until now.
In the early part of my research career and as a research student at the University of Southern California (USC), I concentrated on the design of efficient very-large-scale integration (VLSI) architectures and parallel algorithms, especially for image and signal processing. Such research focused on my development of fast algorithms for solving geometric problems on the Mesh-of-Trees architecture. These techniques have been applied to several other architectures, including bus-based architectures and later on architectures such as the systolic reconfigurable mesh. Thirty years since their inception, these results are still showing their utility in the design of graphics processing unit (GPU) architectures.
Later, as part of my Ph.D., I focused my attention on applying my Mesh-of-Trees results to the area of optical computing. I produced the Optical Model of Computation (OMC) model, through which I was able to show the computational limits and the space-time tradeoffs for replacing electrical wires with free-space optical beams in VLSI chips. Based on the model, I designed several generic electrooptical architectures, including the electrooptical crossbar design that includes a switching speed in the order of nanoseconds. This design was later extended to an architecture called "optical reconfigurable mesh" (ORM). Algorithms designed on ORM have a very fast running time because ORM comprises a reconfigurable mesh in addition to having both a microelectromechanical system (MEMS) and electrooptical interconnectivity. OMC is a well-referenced model that has been shown to have superior performance compared to many other parallel and/or optical models. Based on OMC, the well-known local memory parallel random access memory (PRAM) model was developed. Furthermore, variations of OMC were adopted by the industry in designing MEMS chips.
Soon after I graduated, I took a leading role in starting the heterogeneous computing field. I am the editor of the field's first book, Heterogeneous Computing , and the cofounder of the IEEE Heterogeneous Computing Workshop. The book in conjunction with the workshop shaped the field and paved the path to today's "cloud computing." As one of the first paradigms for executing heterogeneous tasks on heterogeneous systems, I developed the Cluster-M model. Prior models such as PRAM and LogP each had their limitations because they could not handle arbitrary systems or structures with heterogeneous computing nodes and interconnectivity. Cluster-M mapping is still the fastest known algorithm for mapping arbitrary task graphs onto arbitrary system graphs.
For over a decade now, I have been focusing on the bio- and nanoapplications of my work. I am a founding series coeditor of "Nature-Inspired Computing" for John Wiley & Sons and have edited the first book of this series, Bio-inspired and Nanoscale Integrated Computing . This is truly a multidisciplinary topic that required a significant amount of training from several fields. Toward this multidisciplinary field, I have studied various techniques for designing nanoscale computing architecture