Snap Inc.

Jiayao Yu

Jun Huang

Michael David Marr

Michael Ronald Cieslak

Zahra Ferdowsi

Farnaz Azmoodeh

Kai Chen

Chao Pang

Aleksandr Zhang

Finn Parnell

Olamide Valerie Olatunji

Claire Reinert

David Boyle

Sergey Tulyakov

Dhritiman Sagar

Jian Ren

Richard Leider

Kavya Venkata Kota Kopparapu

Benjamin Dodson

Francesc Xavier Drudis Rius

Angus Kong

Fedir Poliakov

Istvan Marton

Cristian Zaloj

Systems, devices, media, and methods are presented for throttling (i.e., adjusting) the workload of an application (e.g., number of task requests) in order to improve processor core usage within a heterogeneous multiprocessor system. When high-performance processing is beneficial to the application, the number of task requests may be increased in order to have high-performance processor cores within the heterogeneous multiprocessor system core processor perform the tasks. On the other hand, when high-performance processing is not beneficial, the number of task requests may be decreased in order to have low-performance processor cores within the heterogeneous multiprocessor system perform the tasks. Processor core usage is monitored, and the number of tasks being performed are adjusted to match the processor core usage to a target processor core usage for functions the application is performing.

Core targeting in heterogeneous multiprocessor systems

Aspects of the present disclosure involve a system comprising a medium storing a program and method for machine-learning based selection of a representative video frame. The program and method provide for receiving a set of video frames; determining a first subset of frames by removing frames outside of an image quality threshold; determining a second subset by removing frames outside of an image stillness threshold; computing feature data for each frame in the second subset; providing, for each frame in the second subset, the feature data to a machine learning model (MLM), the MLM being configured to output a score for each frame in the second subset of frames based on the feature data, the MLM having been trained with a first set of images labeled based on aesthetics, and with a second set of images labeled based on image quality; and selecting a frame based on output scores.

Machine learning-based selection of a representative video frame within a messaging application

Providing branch prediction of APIs to determine the best API for an application to use when performing a function. An API history buffer stores a history of performance data in association with APIs when the function is performed by an application. An API map dynamically stores application performance data in association with APIs. This history data and the API map is used by a circumstance engine and a circumstance engine service to provide a suggested API to an application when it is ready to perform a function.

Branch prediction for APIs

Systems, devices, media, and methods are presented for testing the capability of a mobile device to run a particular feature of an application. Each feature may include a package of code, a call, a state, and a condition; each condition correlated with a specific state. The call is a query that includes instructions to evaluate whether the application will run successfully if a particular feature is running. If the call fails, the feature condition remains Off, resulting in a rollback of the feature. If the call succeeds, the feature condition is changed to On. The system may include an experimentation framework on each mobile device for registering features, executing calls, logging results, and maintaining states and conditions.

Feature rollback for incapable mobile devices

Systems, devices, media, and methods are presented for releasing an application feature in incremental stages while monitoring the application for anomalies. The feature includes a package of code and an action setting. The methods in some implementations include identifying active devices on which the application has been installed, monitoring the application according to a set of metrics, activating the feature by changing its action setting for a first segment of the active devices, pausing the feature if an anomaly is detected among the set of metrics, and generating a repair ticket. As long as no anomaly is detected, the activating step proceeds for subsequent segments of the active devices, iteratively, until the release is completed. A feature rank may be used to process and release a plurality of features in order of priority.

Staged release of updates with anomaly monitoring

Systems, devices, media, and methods are presented for splitting decision trees between server and client. The client of the systems and methods sends a configuration query. The server of the system and method receives the configuration query. The server retrieves Config rule(s) according to the configuration query. Each of the Config rule(s) can be represented by decision tree(s). The server evaluates the decision tree(s). If a definitive True or False cannot be derived from the evaluation using server knowledge, the server prunes the decision tree(s) and returns them to client side for further evaluation.

Split decision trees on client and server

Aspects of the present disclosure involve systems and methods for performing operations comprising providing a messaging application comprising a feature to a client device, the feature being implemented by operations having alternative complexity levels, wherein a first complexity level represents a first amount of device resources consumed by a first set of operations, and wherein a second complexity level represents a second amount of device resources consumed by a second set of operations; determining that the first configuration rule is satisfied by a first property of the client device; and in response to determining that the first configuration rule is satisfied by the first property of the client device, causing the feature to be implemented on the client device by the first set of operations having the first complexity level that consume a greater amount of device resources than the second set of operations having the second complexity level.

Automated scaling of application features based on rules

Systems, devices, media, and methods are presented for assigning configuration parameters to unknown devices by comparing characteristics of unknown devices to characteristics of known devices. Characteristics of an unknown device are compared to characteristics of known devices to identify a similar device having known configuration parameters. The unknown devise is then assigned the configuration parameters of the identified similar device.

Device configuration parameter determination

Systems, devices, media, and methods are presented for determining performance metrics of client devices on a network using benchmark applications. Benchmark applications are distributed to client devices to produce performance metrics for the client devices. Performance metrics of the client devices received from the client devices are grouped to obtain performance metrics for different types of client devices.

Network device performance metrics determination

Aspects of the present disclosure involve systems and methods for performing operations comprising receiving, with a messaging application, user input to access a graphical image modification feature of the messaging application; in response to receiving, causing display of a video; accessing a first configuration rule of a plurality of configuration rules that associates a first device property rule with the graphical image modification feature of the messaging application; determining that the first configuration rule is satisfied by a first property of the client device; and in response to determining that the first configuration rule is satisfied by the first property of the client device, causing display of a first plurality of graphical image modification options each associated with performing a different modification to the video.

Automated graphical image modification scaling based on rules

An example method comprises: receiving, at the server from a first client device, a request for access to a client feature on the first client device; determining, by the server, an applicable rule for the access request, the applicable rule having a plurality of nodes; determining, by the server, device capabilities needed for the determined rule; determining, by the server, nodes that can be executed and nodes that cannot be executed, based on the device capabilities; executing, by the server, nodes that can be executed to reach a partial decision for the applicable rule; pruning the rule to remove executed nodes and generate a pruned rule that includes nodes that cannot be executed; transmitting the pruned rule and partial decision to the device. The pruned rule is executed on the first client device with partial decision to generate a final decision. The client feature is configured based on the decision.

Messaging system with circumstance configuration framework

An example method comprises: receiving, at a server from a first client device, a request for access to a client feature on the first client device; determining, by the server, an applicable rule for the access request, the applicable rule having a plurality of nodes; determining, by the server, device capabilities needed for the determined rule; determining, by the server, nodes that can be executed and nodes that cannot be executed, based on the device capabilities, the nodes that can be executed including device hardware capabilities and the nodes that cannot be executed including real-time device capabilities; executing, by the server nodes that can be executed to reach a partial decision for the applicable rule; pruning the applicable rule to remove executed nodes and generate a pruned rule that includes nodes that cannot be executed; transmitting the pruned rule and partial decision to the first client device.

Growing community of inventors

Venice, CA, United States of America

Jiayao Yu