Juniper Networks, Inc.

Adobe, Inc.

Hui Qu

Shabnam Ghadar

Mu Lin

Elya Shechtman

Jingwan Lu

Ratheesh Kalarot

Joel Abraham Obstfeld

Wei-An Lin

Zi Zhou

David Delano Ward

Baldo Antonio Faieta

Saeid Motiian

Cameron Smith

Shine-Chu Wang

Siavash Khodadadeh

John Thomas Nack

Sven Olsen

Timothy M Converse

Yanfang Zhang

Heng-Hui Debbie Chu

Leyi Zhang

Anjali Agarwal

Anjali Kulkarni

Systems and methods for image processing are provided. Embodiments include identifying an image of a face that includes an artifact in a part of the face. A machine learning model generates an intermediate image based on the original image. The intermediate image depicts the part of the face in a closed position. Then the model generates a corrected image based on the intermediate image. The corrected image depicts the face with the part of the face in an open position and without the artifact.

Repairing irregularities in computer-generated images

Methods, systems, and non-transitory computer readable media are disclosed for intelligently enhancing details in edited images. The disclosed system iteratively updates residual detail latent code for segments in edited images where detail has been lost through the editing process. More particularly, the disclosed system enhances an edited segment in an edited image based on details in a detailed segment of an image. Additionally, the disclosed system may utilize a detail neural network encoder to project the detailed segment and a corresponding segment of the edited image into a residual detail latent code. In some embodiments, the disclosed system generates a refined edited image based on the residual detail latent code and a latent vector of the edited image.

Enhancing detailed segments in latent code-based edited digital images

An item recommendation system receives a set of recommendable items and a request to select, from the set of recommendable items, a contrast group. The item recommendation system selects a contrast group from the set of recommendable items by applying a image modification model to the set of recommendable items. The image modification model includes an item selection model configured to determine an unbiased conversion rate for each item of the set of recommendable items and select a recommended item from the set of recommendable items having a greatest unbiased conversion rate. The image modification model includes a contrast group selection model configured to select, for the recommended item, a contrast group comprising the recommended item and one or more contrast items. The item recommendation system transmits the contrast group responsive to the request.

Semantic-aware initial latent code selection for text-guided image editing and generation

Techniques are described for providing high availability (HA) during an in-service software upgrade (ISSU) of a network device, e.g., a router, by running a pair of virtual machines on each of a primary control unit and a backup control unit within the router. First virtual machines may be run on the primary and backup control units to provide high availability for the router. Upon receiving an ISSU request for the router, second virtual machines may be initialized on the primary and backup control units. If the ISSU of the primary control unit fails such that the first and second virtual machines on the primary control unit are executing different software systems, the virtual machines on the backup control unit may provide high availability for the router. Similarly, the virtual machines on the primary control unit may provide high availability for the router during the ISSU of the backup control unit.

High availability in-service software upgrade using virtual machine instances in dual control units of a network device

Techniques are described for providing high availability during an in-service software upgrade (ISSU) of an appliance within a network device, e.g., a router, by running a pair of virtual machines on each of a primary appliance and a secondary appliance within the router. Examples of the appliances include a routing engine within a router, and a service physical interface card (PIC) within a forwarding engine of a router. An ISSU of the primary appliance may first upgrade the operating system instance of a secondary virtual machine, switch operation from a primary virtual machine to the secondary virtual machine, and then upgrade the operating system instance on the primary virtual machine. During the ISSU of the primary appliance, primary and secondary virtual machines on the secondary appliance provide high availability to the virtual machine on the primary appliance executing the original operating system.

High availability in-service software upgrade using virtual machine instances in dual computing appliances

Techniques are described for providing high availability for a network device, e.g., a router, using full memory replication between a primary virtual machine (VM) and a standby VM running on the network device. In one example, the techniques provide hot-standby high availability for a router by initializing a primary VM and a standby VM on the router at the same time. In another example, the techniques provide high availability for the router by initializing a standby VM on the router upon detecting a failure of the primary VM. In both examples, the primary VM controls the routing functionality and periodically stores a full memory state. Upon a failure of the primary VM, a hypervisor may perform a full memory replication between the primary VM and the standby VM. The standby VM may then take control of the routing functionality at the last stored state of the memory.

High availability using full memory replication between virtual machine instances on a network device

The control plane of a network device comprises a plurality of software processes that manage routing control operations of the device. Through a hypervisor in the control plane, a managing virtual machine controls access to a first virtual machine running a first software system to control a routing communication session between the network device and other network devices. In response to an in-service software upgrade request, the managing virtual machine initializes a second virtual machine. On the second virtual machine, the second software system is loaded. State data maintained by the managing virtual machine can be transferred to the second virtual machine, and the second virtual machine takes control of the routing communication session. During the transfer of control from the first virtual machine to the second virtual machine, techniques of 'non-stop forwarding' and 'graceful restart' can be implemented to minimize the effect the switchover has on the network.

Configuration update on virtual control plane

Techniques described in this disclosure relate to configuration updates, such as performing an in-service software upgrade on a device, using virtual machines. In a routing device, a routing engine utilizes a virtual machine executing on a hypervisor to provide control plane functions. In one example, an in-service software upgrade may be performed between a first virtual machine and a second virtual machine without a managing virtual machine. More specifically, a first virtual machine in the control plane of the router may control the upgrade process, including requesting initialization of the second virtual machine, installing a new software system on the second virtual machine, and replicating state data from the first virtual machine to the second virtual machine. In this example, the first virtual machine may operate as a master virtual machine and the second virtual machine may operate as a slave virtual machine that synchronizes with the master virtual machine.

In-service configuration upgrade using virtual machine instances

Growing community of inventors

Santa Clara, CA, United States of America

Hui Qu