# US 2020/0401522 A1 · Systems and methods for providing content > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2020/0401522 A1 - Document ID: US20200401522A1 - Status: Published application - Kind: A1 - Application number: 15/942,127 - Filed: 2018-03-30 - Published: 2020-12-24 - Earliest claimed: 2018-03-30 - Assignees: Facebook, Inc. - Applicants: Facebook, Inc. - Inventors: Chenyong Xu - Web page: https://patentgrep.com/patent/US20200401522A1 ## CPC - G06F 12/0862: with prefetch - G06F 16/9574: {of access to content, e.g. by caching} - G06N 20/00: Machine learning - G06F 2212/602: Details relating to cache prefetching - G06N 5/022: Knowledge engineering; Knowledge acquisition - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Systems, methods, and non-transitory computer-readable media can determine that a user is interacting with a software application running on a computing device. One or more content items to be prefetched for the software application are identified based on one or more machine learning models. A request to prefetch the one or more content items for the software application is generated. ## Claims ### Claim 1 A computer-implemented method comprising: training, by a computing system, a prefetch machine learning model based on training data received from a plurality of computing devices associated with a plurality of users, wherein the training data includes user characteristics and labeling of previous occurrences of content interaction as examples, wherein the user characteristics include content affinity scores determined based on an affinity machine learning model, a content affinity score being indicative of a user's predicted interest in a content item; determining, by the computing system, that a first user is interacting with a software application running on a computing device; identifying, by the computing system, one or more content items to be prefetched for the software application based on prefetch scores generated by the prefetch machine learning model, a prefetch score being indicative of a likelihood of the first user to view a content item; and generating, by the computing system, a request to prefetch the one or more content items for the software application. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, wherein the prefetch machine learning model is trained based on a set of labels. ### Claim 3 (depends on 2) The computer-implemented method of claim 2, wherein historical instances in which content was prefetched for a particular user and the particular user viewed the prefetched content are positive examples for training the prefetch machine learning model. ### Claim 4 (depends on 2) The computer-implemented method of claim 2, wherein historical instances in which content was prefetched for a particular user and the particular user did not view the prefetched content are negative examples for training the prefetch machine learning model. ### Claim 5 (depends on 2) The computer-implemented method of claim 2, wherein historical instances in which content was not prefetched for a particular user, and the particular user viewed the content are negative examples for training the prefetch machine learning model. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein the identifying the one or more content items to be prefetched further comprises identifying, for each content item of the one or more content items, a portion of the content item to be prefetched based on the prefetch machine learning model. ### Claim 7 (depends on 6) The computer-implemented method of claim 6, wherein, for each content item of the one or more content items, the portion of the content item to be prefetched is determined based on historical user tendencies associated with the user. ### Claim 8 (canceled) ### Claim 9 (depends on 1) The computer-implemented method of claim 1, wherein the one or more content items to be prefetched are identified based on a determination that the one or more content items satisfy a prefetch score threshold. ### Claim 10 (canceled) ### Claim 11 A system comprising: at least one processor; and a memory storing instructions that, when executed by the at least one processor, cause the system to perform a method comprising: training a prefetch machine learning model based on training data received from a plurality of computing devices associated with a plurality of users, wherein the training data includes user characteristics and labeling of previous occurrences of content interaction as examples, wherein the user characteristics include content affinity scores determined based on an affinity machine learning model, a content affinity score being indicative of a user's predicted interest in a content item; determining that a first user is interacting with a software application running on a computing device; identifying one or more content items to be prefetched for the software application based on prefetch scores generated by the prefetch machine learning model, a prefetch score being indicative of a likelihood of the first user to view a content item; and generating a request to prefetch the one or more content items for the software application. ### Claim 12 (depends on 11) The system of claim 11, wherein the prefetch machine learning model is trained based on a set of labels. ### Claim 13 (depends on 12) The system of claim 12, wherein historical instances in which content was prefetched for a particular user and the particular user viewed the prefetched content are positive examples for training the prefetch machine learning model. ### Claim 14 (depends on 12) The system of claim 12, wherein historical instances in which content was prefetched for a particular user and the particular user did not view the prefetched content are negative examples for training the prefetch machine learning model. ### Claim 15 (depends on 12) The system of claim 12, wherein the minimum background time threshold wherein historical instances in which content was not prefetched for a particular user, and the particular user viewed the content are negative examples for training the prefetch machine learning model. ### Claim 16 A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising: training a prefetch machine learning model based on training data received from a plurality of computing devices associated with a plurality of users, wherein the training data includes user characteristics and labeling of previous occurrences of content interaction as examples, wherein the user characteristics include content affinity scores determined based on an affinity machine learning model, a content affinity score being indicative of a user's predicted interest in a content item; determining that a first user is interacting with a software application running on a computing device; identifying one or more content items to be prefetched for the software application based on prefetch scores generated by the prefetch machine learning model, a prefetch score being indicative of a likelihood of the first user to view a content item; and generating a request to prefetch the one or more content items for the software application. ### Claim 17 (depends on 16) The non-transitory computer-readable storage medium of claim 16, wherein the prefetch machine learning model is trained based on a set of labels. ### Claim 18 (depends on 17) The non-transitory computer-readable storage medium of claim 17, wherein historical instances in which content was prefetched for a particular user and the particular user viewed the prefetched content are positive examples for training the prefetch machine learning model. ### Claim 19 (depends on 17) The non-transitory computer-readable storage medium of claim 17, wherein historical instances in which content was prefetched for a particular user and the particular user did not view the prefetched content are negative examples for training the prefetch machine learning model. ### Claim 20 (depends on 17) The non-transitory computer-readable storage medium of claim 17, wherein historical instances in which content was not prefetched for a particular user, and the particular user viewed the content are negative examples for training the prefetch machine learning model. ### Claim 21 (depends on 11) The system of claim 11, wherein the one or more content items to be prefetched are identified based on a determination that the one or more content items satisfy a prefetch score threshold. ### Claim 22 (depends on 16) The non-transitory computer-readable storage medium of claim 16, wherein the one or more content items to be prefetched are identified based on a determination that the one or more content items satisfy a prefetch score threshold.