Existing sequential feature-based registration algorithms involving search typically either select features randomly (eg. the RANSAC[8] approach) or assume a predefined, intuitive...
In this paper, we describe a methodology to estimate the geographic coverage of the web without the need for secondary knowledge or complex geo-tagging. This is achieved by random...
Robert Pasley, Paul Clough, Ross S. Purves, Floria...
We introduce a new approach to analyzing click logs by examining both the documents that are clicked and those that are bypassed--documents returned higher in the ordering of the ...
Atish Das Sarma, Sreenivas Gollapudi, Samuel Ieong
We address the task of learning rankings of documents from search engine logs of user behavior. Previous work on this problem has relied on passively collected clickthrough data. ...
In this paper we study the possibility of proving the existence of one-way functions based on average case hardness. It is well-known that if there exists a polynomial-time sample...