Microsoft Technology Licensing, LLC

Ted Tomlinson

Daniel Francis

Shi Yan

Junzhe Miao

Xiaochuan Zhou

Paul D Bergeron

Nikko David Bautista

Gurbir Singh

The disclosed embodiments provide a system for processing data. During operation, the system extracts text windows of varying length from text in one or more content items associated with an entity. Next, the system applies a machine learning model to features for the text windows to produce scores representing the likelihoods that the text windows contain addresses. The system then identifies, based on the scores and validation rules applied to the text windows, one of the text windows as an address for the entity. Finally, the system stores the selected text window as the address for the entity.

Identifying and extracting addresses within content

Apparatuses, computer readable medium, and methods are disclosed for dynamic sampling based on talent pool size. The method of dynamic sampling based on talent pool size may include determining, by at least one hardware processor, a talent pool based on a query from a user, where the talent pool comprises members and the members comprise attributes. The method further includes determining based on a size of the talent pool a sampling size number, where the attributes are represented by probability distribution functions. The method may further include determining an aggregate distribution of the attributes by combining the sampling size number of samples from each of the attributes and determining a range of a value of the aggregate distribution of the attributes. The method may further include causing the range of the value of the aggregate distribution of the attributes to be displayed to the user.

Dynamic sampling based on talent pool size

Growing community of inventors

Oakland, CA, United States of America

Ted Tomlinson