Florida International University

Facebook, Inc.

Weitong Liu

Tao Li

Naphtali David Rishe

Mingjin Zhang

Jahkell Lazarre

Abhishek Bapna

Eric Hsin-Chun Huang

Shuo Li

Menghan Li

The present disclosure is directed toward systems and methods for dynamically providing digital content to client devices at different insertion points of a digital video based on predicted total value of inserting the digital content and predicted engagement loss of inserting the digital content. For example, system and methods described herein determine that an insertion point is approaching in an actively playing digital video. In response, systems and methods identify digital content for insertion at the insertion point. In one or more embodiments, the described systems and methods insert the identified digital content by comparing the predicted total value of inserting the digital content at the insertion point of the digital video and a predicted engagement loss associated with inserting the digital content at the insertion point of the digital video.

Dynamically providing digital content to client devices by analyzing insertion points within a digital video

Devices and methods for providing Voronoi diagram-based progressive continuous range query (CRQ) computation for location-based services are provided. A method can include indexing all interest points using a Voronoi R-tree data structure in a server-side computing device. When a client mobile device sends a CRQ request, the server-side device receives a path of travel and builds a list to store interest points and a min-heap to store split points, iteratively retrieving split points from the min-heap, and reporting the CRQ result progressively until a termination condition is reached.

Voronoi diagram-based algorithm for efficient progressive continuous range query for moving objects

Methods and systems for computing continuous k nearest neighbor (CkNN) queries in location based services for moving objects, to produce ordered kNN query results in a continuous and progressive manner, are provided. One method comprises receiving a continuous k nearest neighbor query, computing the initial set comprised of the k interest points nearest to the reference point and a set of remaining nodes stored in a distance-priority queue, generating and storing split points in a min-heap, iteratively moving current reference point to the nearest split point, swapping interest points, updating the corresponding split points in the min-heap, and reporting the kNN result progressively until a termination condition is reached.

Efficient progressive continuous k-nearest neighbor query algorithm for moving objects with a tree-like index

Methods and apparatuses for providing a k-nearest neighbor for location based services are provided. A method can include querying a database to detect a plurality of interest points within a predetermined distance of the user device using a kNN algorithm, organizing the interest points within a Voronoi tree, and continuously return a position specific result of relevant interest points.

Voronoi diagram-based algorithm for efficient progressive continuous k-nearest neighbor query for moving objects

Methods for progressive continuous range query (PCRQ) are provided. A method can include using branch-and-bound to index interest points with a tree-index and generating a nearest enter split-point for a root node in the tree-index and adding to min-heap. Next, whether min-heap has more elements and whether a next split-point in min-heap is closer than a destination can be determined. Whether a query point has reached a split-point can be investigated followed by retrieving an entry that has generated the split-point. The split-point can be then be removed from min-heap.

Growing community of inventors

Redwood City, CA, United States of America

Weitong Liu