BRENT HECHT
Microsoft Research profile
Google Scholar profile
LinkedIn
Director of Applied Science
Experiences and Devices
Microsoft
brent(dot)hecht(at)microsoft.com
2240 Campus Rd., Evanston, IL
Lab:
blah blah blah
(Coming Soon to GitHub)
Many applications of popular semantic relatedness measures (e.g. Explicit Semantic Analysis, WikiRelate) frequently calculate the relatedness between two named entities or one named entity and a common noun. However, most existing semantic relatedness human gold standard datasets are focused on pairs of common nouns (e.g. "boy" and "lad"). Atlasify240, which was developed for our Atlasify project (pictured at left), helps remedy this situation. It contains a large number of named entities such as countries, animals, historical figures, etc.
For more information, please see our 2012 SIGIR paper. If you use this dataset, please cite this paper.
In the course of making various presentations, I've developed a number of icons. Since most of them were derived or inspired by those on The Noun Project, I'm working to convert them into vector format and make them available under a CC BY 3.0 license there.