US 11,899,733 B2Granted

Method and system for activity prediction, prefetching and preloading of computer assets by a client-device

AssigneeGoogle Inc.

InventorsMichael Shalai, Joseph Catalano, Bo Lin, Dustin Zelle, Rami Al-Rfou

Life of this patent

Estimating the term…

Loading patent…

Abstract

A solution arranged to build or train a machine learning model (ML model) that can be uploaded to a server arranged to deploy the ML model to communicating devices. The ML model builder can build the ML model and a ML production pipeline. The ML production pipeline can train the ML model, convert the ML model to a web browser compatible format, and upload the converted ML model to the server. The ML model can receive as input a sequence of prior activities on one communicating device in the communicating devices, analyze the sequence of prior activities on the communicating device, predict a next activity on the communicating device based on the analysis of the sequence of prior activities, preemptively search a computer network based on the predicted next activity to find a computer asset, and preload the found computer asset to a storage in the communicating device.

Claims20

Dependent claims

A computer implemented method for activity prediction, the method comprising:

building, by a MLM builder associated with the computing device, a machine learning model using activity log data received by the MLM builder, wherein the MLM builder extracts activity information and groups the extracted activity information by session;
training, by the computing device, the machine learning model by applying training activity datasets to the machine learning model, wherein the machine learning model is continuously trained to serve an up-to-date machine learning model;
converting, by the computing device, the machine learning model into a web browser compatible format;
and
uploading, by the computing device, the machine learning model to a server that is arranged to deploy the machine learning model to a plurality of communicating devices, wherein the machine learning model is arranged to:
receive, by a machine learning model, as input a sequence of one or more prior activities on one communicating device of a plurality of communicating devices, the sequence comprising the one or more prior activities during a session in chronological order, the one or more prior activities comprising at least one of page transitions, button clicks, or remote procedure calls (RPCs);
analyze, by the machine learning model, the sequence of one or more prior activities on the one communicating device, wherein the analyzing includes calculating relevance weights for search results, wherein the relevance weights are indicative of a relevance to prediction data, and identifying the results having the greatest relevance to the prediction data;
predict, by the machine learning model, a next activity on the one communicating device based on the analysis of the sequence of one or more prior activities;
search a computer network based on the predicted next activity to find a computer asset, wherein the computer asset comprises a next application;
and
preload the computer asset to a storage in the one communicating device.

8 more elements

A computing system, comprising:

one or more processors;
and
one or more non-transitory computer-readable media storing instructions that are executable to cause the one or more processors to perform operations, the operations comprising:
building, by a MLM builder associated with the computing device, a machine learning model using activity log data received by the MLM builder, wherein the MLM builder extracts activity information and groups the extracted activity information by session;
training a machine learning model by applying training activity datasets to the machine learning model, wherein the machine learning model is continuously trained to serve an up-to-date machine learning model;
converting the machine learning model to a web browser compatible format;
and
uploading the converted machine learning model to a server arranged to deploy the machine learning model to a plurality of communicating devices,
wherein the machine learning model is arranged to:
receive as input a sequence of one or more prior activities on one communicating device in the plurality of communicating devices, the sequence comprising the one or more prior activities during a session in chronological order, the one or more prior activities comprising at least one of page transitions, button clicks, or remote procedure calls (RPCs);
analyze the sequence of one or more prior activities on the one communicating device, wherein the analyzing includes calculating relevance weights for search results, wherein the relevance weights are indicative of a relevance to prediction data, and identifying the results having the greatest relevance to the prediction data;
predict a next activity on the one communicating device based on the analysis of the sequence of one or more prior activities;
search a computer network based on the predicted next activity to find a computer asset, wherein the computer asset comprises a next application;
and
preload the computer asset to a storage in the one communicating device.

12 more elements

A non-transitory computer-readable storage medium containing computer executable instructions that, when executed by a computing device, cause the computing device to perform operations, the operations comprising:

building, by a MLM builder associated with the computing device, a machine learning model using activity log data received by the MLM builder, wherein the MLM builder extracts activity information and groups the extracted activity information by session;
training, by the computing device, the machine learning model by applying training activity datasets to the machine learning model, wherein the machine learning model is continuously trained to serve an up-to-date machine learning model;
converting, by the computing device, the machine learning model into a web browser compatible format;
and
uploading, by the computing device, the machine learning model to a server that is arranged to deploy the machine learning model to a plurality of communicating devices,
wherein the machine learning model is arranged to:
receive as input a sequence of one or more prior activities on one communicating device in the plurality of communicating devices, the sequence comprising the one or more prior activities during a session in chronological order, the one or more prior activities comprising at least one of page transitions, button clicks, or remote procedure calls (RPCs);
analyze the sequence of one or more prior activities on the one communicating device, wherein the analyzing includes calculating relevance weights for search results, wherein the relevance weights are indicative of a relevance to prediction data, and identifying the results having the greatest relevance to the prediction data;
predict a next activity on the one communicating device based on the analysis of the sequence of one or more prior activities;
search a computer network based on the predicted next activity to find a computer asset, wherein the computer asset comprises a next application;
and
preload the computer asset to a storage in the one communicating device.

9 more elements

Dependencies are inferred from the claim text.