Abstract: Various implementations disclosed herein include devices, systems, and methods for pose estimation using one point correspondence, one line correspondence, and a directional measurement. In various implementations, a device includes a non-transitory memory and one or more processors coupled with the non-transitory memory. In some implementations, a method includes obtaining an image corresponding to a physical environment. A first correspondence between a first set of pixels in the image and a spatial point in the physical environment is determined. A second correspondence between a second set of pixels in the image and a spatial line in the physical environment is determined. Pose information is generated as a function of the first correspondence, the second correspondence, and a directional measurement.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating output sequences using auto-regressive decoder neural networks. In particular, during generation, adaptive early exiting is used to reduce the time required to generate the output sequence.

Abstract: The disclosure relates to integrin binding peptides, pharmaceutical compositions comprising the peptides and to uses thereof as therapeutic, diagnostic, imaging and targeting agents.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating output sequences using auto-regressive decoder neural networks. In particular, during generation, adaptive early exiting is used to reduce the time required to generate the output sequence.

Abstract: Some implementations involve a process that identifies a subset of points in an image and creates descriptors for these points. The detection and descriptor process may use one or more neural networks. In some implementations, the process includes a neural network that uses one or more fixed (e.g., weight independent) neural network layers to perform certain functions that can be performed more accurately and/or efficiently than via non-fixed (e.g., weight-based) layers. In some implementations, for example, a neural network includes a layer that determines orientation formulaically within the neural network. Such orientations may be determined convolutionally (e.g., using sliding patches) but are not determined based on internal node weights within the layer that were determined during the training of the neural network.

Abstract: A system and a method for producing ethylene glycol are disclosed. Alkylene oxide and water are flowed into a first reactor unit and subjecting the alkylene oxide and water, in the first reactor unit, to first reaction conditions such that an effluent of the first reactor unit comprises an alkylene glycol, unreacted alkylene oxide, and unreacted water. At least a portion of the unreacted alkylene oxide may be routed to a second reaction unit and subjected to reaction conditions sufficient to produce additional alkylene glycol, wherein the second reactor unit is a reactive distillation column.

Abstract: In one implementation, a method of estimating the heading of a device is performed by the device including a processor, non-transitory memory, and an image sensor. The method includes determining a geographic location of the device. The method includes capturing, using the image sensor, an image at the geographic location. The method includes detecting one or more lines within the image. The method includes determining a heading of the device based on the one or more lines and the geographic location.

Abstract: In one implementation, a method of estimating the heading of a device is performed by the device including a processor, non-transitory memory, and an image sensor. The method includes determining a geographic location of the device. The method includes capturing, using the image sensor, an image at the geographic location. The method includes detecting one or more lines within the image. The method includes determining a heading of the device based on the one or more lines and the geographic location.

Abstract: A composition comprising a liposome comprising a first phospholipid comprising a C14-C19:0 fatty acid and a C3-C15:0 fatty acid with a c-terminal carboxyl or a carboxaldehyde group; a second phospholipid comprising a C14-C19:0 fatty acid and a C14-C19:0 fatty acid and a sterol. The first phospholipid is preferably a phosphatidylcholine. The second phospholipid is preferably HSPC and the sterol is preferably cholesterol. In some embodiments, the molar ratio of the respective ingredients is 2-3:5-6:2-3; preferably 2:6:2; or 3:5:2, respectively. The invention also relates to the use of this composition in the targeting of M2 macrophages.

Abstract: The disclosure relates to integrin binding peptides, pharmaceutical compositions comprising the peptides and to uses thereof as therapeutic, diagnostic, imaging and targeting agents.

Abstract: A packet forwarding network may include spine and leaf switches that forward network traffic between end hosts. The packet forwarding network may be implemented on multiple network racks in a rack-based system. A controller may control the underlying spine and leaf switches to form on-premise virtual private cloud (VPC) resources. In particular, the controller may form enterprise VPC (EVPC) tenants, each having a virtual router that performs routing between different segments within the corresponding EVPC tenant. The different segments may separately include web, application, and database servers, as end hosts. The controller may form a system VPC tenant having a virtual system router that performs routing between different EVPC tenants. A segment in an internal VPC tenant formed by the controller and/or an external VPC tenant formed by the controller may provide external network access for one or more of the EVPC tenants.

Abstract: Provided are machine-learned attention models that feature omnidirectional processing, example implementations of which can be referred to as Omnidirectional Representations from Transformers (OMNINET). In example models described in the present disclosure, instead of maintaining a strictly horizontal receptive field, each token is allowed to attend to all tokens in some or all of the other tokens across the entire network.

Abstract: Various implementations disclosed herein include devices, systems, and methods for pose estimation using one point correspondence, one line correspondence, and a directional measurement. In various implementations, a device includes a non-transitory memory and one or more processors coupled with the non-transitory memory. In some implementations, a method includes obtaining an image corresponding to a physical environment. A first correspondence between a first set of pixels in the image and a spatial point in the physical environment is determined. A second correspondence between a second set of pixels in the image and a spatial line in the physical environment is determined. Pose information is generated as a function of the first correspondence, the second correspondence, and a directional measurement.

Abstract: Various implementations disclosed herein include devices, systems, and methods for pose estimation using one point correspondence, one line correspondence, and a directional measurement. In various implementations, a device includes a non-transitory memory and one or more processors coupled with the non-transitory memory. In some implementations, a method includes obtaining an image corresponding to a physical environment. A first correspondence between a first set of pixels in the image and a spatial point in the physical environment is determined. A second correspondence between a second set of pixels in the image and a spatial line in the physical environment is determined. Pose information is generated as a function of the first correspondence, the second correspondence, and a directional measurement.

Abstract: Systems, methods, and computer-readable media are described for creating custom user interface (“UI”) controls. A networked computing service may determine that an action performed by a user of the service is likely to be repeated, based on previous actions performed by the user and/or other users, parameters of the action including characteristics of an item associated with the action, and other criteria. The networked computing service may generate, or recommend generation of, a custom UI control that automates performance of the action. The networked computing service may automatically determine parameters of the action, and may determine alternative parameters and/or alternative actions if the action cannot be performed using the determined parameters. Custom UI controls may be associated with various computing devices and/or other electronic devices, and categories of custom UI controls may be determined and displayed.

Abstract: Filename-based malware pre-scanning is described herein. A method as described herein can include obtaining, by a device operatively coupled to a processor, a malware scan request for a first file in a directory of a file system, the first file having a first filename belonging to a filename sequence; appending, by the device, the first file to a first malware scan queue; and appending, by the device, respective second files in the directory to a second malware scan queue that is distinct from the first malware scan queue, wherein the respective second files are distinct from the first file and have respective second filenames belonging to the filename sequence.

Abstract: In accordance with some embodiments, a method is performed at a device with one or more processors, non-transitory memory, and an image sensor. The method includes synthesizing, from image data that corresponds to a physical perspective of an object relative to the device, a surface mask characterizing a portion of a surface of the object based on a grid arranged on the object, wherein the grid is represented in the image data. The method includes generating a virtual perspective of a model associated with the object having a surface oriented to match the surface mask within a predefined confidence threshold.

Abstract: Filename-based malware pre-scanning is described herein. A method as described herein can include establishing sequential malware scanning for respective files in a first directory of a file system according to filenames of the respective files; comparing an elapsed time since a previous malware scan for a first file of the respective files in the first directory having a first filename in a filename sequence to a scan age threshold associated with a scheduled malware scan task for the first directory; and, in response to the elapsed time since the previous malware scan for the first file being determined to be less than the scan age threshold, removing the scheduled malware scan task with respect to the first file and at least one second file of the respective files in the first directory having respective second filenames in the filename sequence.

Abstract: Efficient configuration replication using a configuration change log is described herein. A method as described herein can include identifying, by a device operatively coupled to a processor, modified configurations from among a group of configurations associated with a first computing cluster; fetching, by the device, the modified configurations from the first computing cluster; and replicating, by the device, the modified configurations from the first computing cluster to a second computing cluster that is distinct from the first computing cluster and omitting replication of one or more unmodified configurations from among the group of configurations associated with the first computing cluster.

Abstract: Priority scanning of files written by malicious users in a data storage system is described herein. A data storage system as described herein can include a user lookup component that obtains identities of users that have made at least one modification to a first file stored on the data storage system, resulting in a set of modifying users; a comparison component that compares respective modifying users of the set of modifying users to respective malicious users of a set of malicious users; and a scan priority component that, in response to the comparison component identifying at least one match between a modifying user of the set of modifying users and a malicious user of the set of malicious users, assigns a first scan priority to the first file that is higher than a second scan priority assigned to a second, different file stored on the data storage system.