Search

A new report from Zhengzhou University in China discusses the need for large-scale constrained multi-objective optimization in personalized medicine. The researchers propose a new benchmark for testing algorithms in this field, taking into account realistic features such as mixed linkages between variables and varying numbers of constraint functions. They also introduce a bidirectional sampling strategy to improve algorithm performance in large-scale search spaces with constraints. The proposed algorithm is shown to be effective in solving personalized drug target recognition problems with over 2000 decision variables. [Extracted from the article]

A technique of random choice is illustrated by application to the problem of assigning students to a fixed schedule of courses. Using the technique it is possible to reduce or eliminate difficulties that result when a popular section is filled and closed before oil students requesting and requiring it have been scheduled. The effectiveness of automatic scheduling is retained without loss of the students' privilege of picking favorite instructors. [ABSTRACT FROM AUTHOR]

The article reports on the development of an algorithm to speed up magnetic resonance imaging (MRI) process by associate professors Elfar Adalsteinsson and Vivek Goyal in 2011. The algorithm uses information derived from the first contrast scan to aid the production of subsequent images. In addition, the algorithm uses the first scan to portend the probable position of the boundaries between various types of tissue in the subsequent contrast scans.

The article discusses the human computable passwords being worked out by the Carnegie Mellon University computer science professor Manuel Blum, who won the Turing Award in 1995. Topics discussed include the use of an algorithm and private key for the creation and re-creation of unique passwords, the alternate letter and number provided by the algorithm in the name of the website, and the work needed for selecting and learning the key and algorithm.

The article profiles Anna Lysyanskaya, an assistant professor of computer science at Brown University and one of the recipients of TR35 Innovator of the Year. Lysyanskaya has developed a practical way for people to securely log in to Web sites without providing any identifying information. She, in collaboration with Jan Camenisch of the IBM Zürich Research Laboratory, developed algorithms that make both generating and testing credentials much more efficient. The algorithms will be incorporated by IBM into its Idemix anonymous-credential systems.