ITR/SY: Combinatorial Optimization Algorithms for Informaion Access (Fundamental IT Models)

点击翻译按钮获取中文摘要

Proposal No: 0113371ITR/SY: Combinatorial Optimization Algorithms for Information Access (Fundamental IT Models)PI: Eva TardosThe enormous growth in the amount of information online makes it vitalto automate the job of searching and organizing large data collectionswhile maintaining high accuracy. The research will consider twoseemingly unrelated tasks:. Searching web pages, and. Analyzing the content of digital pictures and movies.Despite their obvious great importance, and the large efforts invested,the currently available solutions for these tasks are inadequate. Theproposed research will address a difficult mathematical problem thatunderlies both of these tasks, the classification problem with pair-wiseconstraints. The traditional classification problems consist of a set of objects tobe classified, and a set of labels (the classes). Classifying the topicsof documents on the Web, or individual pixels or regions of an image aretwo examples of this general problem. The proposed research will buildon the PIs recent successes in applying powerful combinatorialoptimization techniques to develop algorithms for classificationproblems with pair-wise relationships. Pair-wise constraints cansignificantly enhance classification, by modeling, e.g., the relationsof physically close objects in an image, or relations implicit in thehyperlink structure of the Web. The outcome of this research will be toprovide powerful new tools for two important tasks in searching andorganizing large data collections.