September 1992 The Department of Computer Science at the University of Rochester offers an intense, research-oriented program leading to the degree of Doctor of Philosophy, with particular emphasis currently being placed on the areas of parallel systems, computer vision and robotics, knowledge representation and natural language understanding, and the theory of computation. There is no professional Master's program. The Department is currently in the process of establishing a selective undergraduate major. The Department is tight-knit and concentrates on research and its Ph.D. program. There are currently thirteen full-time faculty and about fifty graduate students. There are also a handful of affiliated faculty from such departments as Philosophy and Electrical Engineering. The focused research interests reflect our desire to achieve excellence in a core of important issues, rather than to try to cover all areas. At the same time, our Department is aggressive and effective in its collaboration with other departments within the University. We have been successful in maintaining research excellence and economic vitality while remaining small enough to support an unusually friendly, open atmosphere. The Department emphasizes intellectual cooperation, top-notch facilities, high leverage research issues, and a faculty committed to students and research (not to consulting or to their own companies). In 1975 Rochester was the first university to obtain Xerox Alto workstations on an Ethernet; in 1985 the Department acquired a 128-node BBN Butterfly Parallel Processor, for several years the largest shared-memory multiprocessor in the world; on a continuing basis our research and funding structure encourages cooperation between theory, systems, and artificial intelligence, and between our own researchers and those of other departments. The Department has always been well-supported by the University and funding agencies. To students this means the best facilities, stipends that continue indefinitely as long as the student remains in good standing, the possibility of summer employment in the Department, and support to cover expenses of attending conferences, workshops, and institutes. The cooperative structure of the Department means that students are free to choose their fields and supervisors on the basis of intellectual attraction, not on the basis of funding. It also means that students play an important role in policy and decision-making, serving on almost every department committee, including those devoted to admissions, faculty recruiting, and the content of the Ph.D. comprehensive exams. Facilities In 1987 the Department moved into the top two floors of a newly-constructed building providing spacious offices for faculty, staff, and all students in the graduate program. There is ample lab space for public workstations and specialized research, a lounge, kitchen, and several class/meeting rooms. The Computer Engineering group in the College of Engineering and Applied Science occupies the two floors below, and the Chester Carlson Science and Engineering Library, which brings together several previous campus libraries, occupies the remainder of the building and makes library research efficient. The Department maintains a high-speed local network of computers for the research use of its faculty, staff, and students. As of summer 1992 the computers on the network include about forty Sun 4 workstations, ten Sun 3 workstations, five Sun file servers, eight large-screen Macintosh II PCs, miscellaneous other workstations, two IBM 8CE multiprocessors, four BBN Butterfly multiprocessors, and an 8-processor top-of-the-line Silicon Graphics multiprocessor. The Silicon Graphics machine is scheduled to be upgraded to R4000 processors this fall. The Computer Vision and Robotics Laboratory is equipped with a Unimation Puma robot arm, a Utah four-fingered dextrous hand, a custom-designed binocular "robot head," and numerous DataCube real-time image processing boards. A state-of-the-art speech lab is under construction. Each of the Department's faculty, staff, and student offices is equipped with at least one workstation, and many additional workstations are maintained in public areas. A high-speed connection to the University's campus network provides excellent access to the NYSERNET regional network, NSFNET, and the Internet. The People The students at Rochester are an elite group. With rare exceptions they are all in the Ph.D. program. They come from all over the world, with the USA and Canada having the strongest representations. They tend to rank in the top 5% on GRE scores, grade-point average, and recommendation letters (the median three-score GRE total is about 2150). In recent years, we have had some four hundred applications per year, and have made approximately twenty offers to yield an entering class of ten or twelve. Students are encouraged to apply for competitive fellowships and are supported in such applications by the faculty. They have received awards from the NSF, AT&T, ONR, DARPA, Eastman Kodak, IBM, and the Ford Foundation, as well as several foreign governments. Students are encouraged to do independent research and to publish and present it in the most influential places possible. Faculty are carefully selected to complement the existing interests and talents of the Department. Since there is no "service" teaching, and since all offers are tenure-track, all faculty are here for the same purpose -- to do and supervise important work. This commitment makes a difference internally and externally; currently on the faculty we have two NSF Presidential Young Investigators, an Office of Naval Research Young Investigator and the recipient of an IBM Faculty Department Award. Postdoctoral research fellows are a factor of increasing importance and contribute expertise in a number of areas. Our administrative and secretarial staff (four full time and several part time) are highly experienced and able to deal effectively with department, university, and external business. Our five full-time members of the technical staff (and several part time staff) are expert professionals who are often enrolled in graduate courses and involved directly in departmental research efforts. The Systems Group The major focus of systems research at the University of Rochester is parallel and distributed computing. Our collection of Silicon Graphics, IBM 8CE, and BBN Butterfly multiprocessors provides an outstanding environment for work in large-scale tightly-coupled systems. Our network of Sun-4 workstations provides an excellent (though less distinctive) environment for work in distributed systems. Our state-of-the art robotics laboratory, comprised of the various multiprocessors, a pipelined image processing computer, cameras, monitors, robot arm, and dextrous robot hand, provides a unique opportunity for work in real-time reactive systems. An example of cooperative work between the systems and vision groups can be found in the 1990 Checkers project, described (among other places) in the February 1992 issue of the IEEE's Computer magazine. Using vision lab equipment, our experimental Psyche parallel operating system, and a team of programmers drawn from the systems and vision groups, we put together a game-playing robot that replaces the human being on one side of a dime-store checkers set -- watching human moves, plotting strategy, and making moves of its own with the robot arm. The principal innovation of the project was its use of several very different models of parallelism -- data parallel image processing, distributed alpha-beta search, and parallel symbolic motion planning. The current systems faculty consists of Tom LeBlanc (chairman), Michael Scott, and Rob Fowler. Additional systems hires are expected in the spring of 1993. Much of the current research is directed at the increasingly important issue of locality -- how to maximize the proximity of threads and data in large-scale parallel systems, via language, runtime, kernel, and hardware-level mechanisms. The Ph.D. Program The Ph.D. program includes courses, examinations, projects, research and teaching assistantships, and a doctoral dissertation. There is only one required course, which serves as an introduction to the Department, to graduate research, and to fellow students and the faculty. Each of the department's research areas (AI, Systems, and Theory) provides a basic syllabus from which students may choose a course of study appropriate to their interests and backgrounds. Several higher-level seminars are also offered each semester, providing more senior students with an opportunity to pursue research topics for classroom credit. Courses in other departments are often of direct or indirect interest as well. There are no formal summer graduate courses. The graduate program encompasses a steady progression from classroom learning to technical competence and independent research. The first year culminates in a Comprehensive Exam in AI, Systems, and Theory, designed to measure breadth. Students typically spend their first year assisting with department research and taking basic courses. The second year serves as a transition into self-directed work; by the end of year two each student is expected to pass an Area Exam assessing depth in one of the Department's three main areas, and to write a technical paper demonstrating critical independent though in some subject within that area. By the end of the third year, each student is expected to find a dissertation advisor and topic, present conference-quality preliminary results, and defend a plan of research for completion of the doctoral dissertation. Satisfactory progress in the program is determined by the full range of a student's work, including performance on the Comprehensive Exam and programming problem, course grades, department service, and research productivity. Research work and research-related courses occupy much of a student's time beyond the first year, and essentially all of it after the Area Exam. After approval of a dissertation topic, each student meets twice yearly with his or her thesis committee, to keep them appraised of progress and to get formal feedback and advice. Ph.D. students in the Department are either supported as research assistants or are fellowship recipients. Each student is also required to serve as a teaching assistant for two or three semesters during his or her course of study. The possibility of assuming primary teaching responsibilities for undergraduate or graduate courses also exists, should a student actively desire such an opportunity. Five years is the customary time from entry to the Ph.D. Financial support continues indefinitely as long as the faculty agrees that the student is making adequate progress. The University The University of Rochester is a private research university located on the banks of the Genesee River in Rochester, New York. It has about 1200 faculty members, 4900 undergraduate students, and 2700 graduate students. It comprises 7 schools and colleges, including the Eastman School of Music and the William E. Simon Graduate School of Business Administration. The Rochester metropolitan area has a population of just over one million people. Its industrial base is principally "high-tech", with a heavy emphasis on optics. Major local employers include Kodak, Xerox, and Bausch and Lomb. Cultural attractions include a symphony orchestra, many professional and semi-professional theatre and dance companies, and a wide variety of outstanding museums. The surrounding Finger Lakes region of Western New York is reknown for its natural beauty. For further information Admissions applications and information can be obtained by sending e-mail to admissions@cs.rochester.edu, or by writing to Graduate Admissions, Computer Science Department, University of Rochester, Rochester, NY 14627-0226. Requests for technical information on the systems group will be forwarded to the current systems member of the graduate admissions committee.