We present a pairwise learning-to-rank approach to machine translation evaluation that learns to differentiate better from worse translations in the context of a given reference. We integrate several layers of linguistic information encapsulated in tree-based structures, making use of both the reference and the system output simultaneously, thus bringing our ranking closer to how humans evaluat...

Sixty patients with advanced breast cancer unresponsive to tamoxifen have been randomised to receive four course of mitozantrone, 14 mg m2 (n = 30) intravenously every 3 weeks (9 weeks total) or megesterol acetate, 160 mg bd (n = 30). One in three patients (11 from each group) had substantial disease control for a minimum period of 6 months i.e., lack of progression; seven patients (23%) showed...

I use the example of the 2000 US Presidential election to show that political controversies with technical underpinnings are not resolved by technical means. Then, drawing from examples such as climate change, genetically modified foods, and nuclear waste disposal, I explore the idea that scientific inquiry is inherently and unavoidably subject to becoming politicized in environmental controver...

OBJECTIVE The prevalence of childhood obesity more than doubled in the period from 1961 to 2001. We replicated a 1961 study of stigma in childhood obesity to see what effect this increased prevalence has had on this stigma. RESEARCH METHODS AND PROCEDURES Participants included 458 5th- and 6th-grade children attending upper-middle and lower-middle income U.S. public schools. Children ranked s...

PURPOSE The aim of this systematic review was to determine which dowel (post) and core system is the most successful when used in vivo to restore endodontically treated teeth. MATERIALS AND METHODS A MEDLINE, a Cochrane, and an EMBASE search (three specified searches) were conducted to identify randomized (RCT) and nonrandomized controlled clinical trials (CCT), cohort (CS), and case control ...