Google Inc.

Contextlogic, Inc.

Piotr Szulczewski

Jack Z Xie

Danny S Zhang

Yuli Ye

Tarek Fahmy

Sheng Zhang

A method for ranking items may include determining a purchaseability score for each item of a specified type based on a number of impressions of the respective item provided to users during a time period, a quantity of the respective item purchased by the users, and a quantity of the respective item added to shopping carts of the users. The method may further include ranking the items based on the purchaseability scores, generating item feed data indicating an ordering of the items based on the rankings, and providing the item feed data to a user device. The user's device may display a scrollable feed of item panels corresponding to the items, with the item panels ordered according to the ordering of the corresponding items.

Systems and techniques for ranking items

Techniques to provide contextual content are disclosed. In various embodiments, a set of topics inferred programmatically from a corpus of raw content data is obtained. An input is received. A topic with which the input is associated is determined, based at least in part on the input and the set of topics. A contextual content associated with the topic is provided.

Topic inference based contextual content

Systems and techniques for ranking items on a scrollable feed using a hybrid user interface

One or more first groups of one or more first content items are identified. A first score is determined for each of the one or more first groups. A second group of one or more second content items is identified. At least one of the one or more second groups correspond to at least one of the one or more first groups. A second score is determined for each of the one or more second groups. A third score is determined for the first groups and the corresponding second group based on the first scores and second scores. An adjustment is made to a ranking of the one or more first content items based on the third scores.

Advertisement topic diversification and ranking

Content of a first type, such as web pages, are identified in response to a query. One or more correlated entities are identified based on the content of the first type. Content of a second type, such as advertisements, are identified based on the one or more of the correlated entities.

Growing community of inventors

San Francisco, CA, United States of America

Piotr Szulczewski