Abstract: A method for encoding a sequence of packets includes receiving the sequence of packets, generating a parity packet for a first group of packets within the sequence of packets, and transmitting the first group of packets and the parity packet. The parity packet is generated by performing an exclusive OR (XOR) operation over a plurality of packets in the first group of packets and at least one packet in a second group of packets. The second group is separated from the first group in the sequence by one or more packets.

Abstract: Techniques for tracking compute capacity of a scalable application service platform to perform dynamic bandwidth allocation for data flows associated with applications hosted by the service platform are disclosed. Some of the techniques may include allocating a first amount of bandwidth of a physical underlay of a network for data flows associated with an application. The techniques may also include receiving, from a scalable application service hosting the application, an indication of an amount of computing resources of the scalable application service that are allocated to host the application. Based at least in part on the indications, a second amount of bandwidth of the physical underlay to allocate for the data flows may be determined. The techniques may also include allocating the second amount of bandwidth of the physical underlay of the network for the data flows associated with the application.

Abstract: Techniques for tracking compute capacity of a scalable application service platform to perform dynamic bandwidth allocation for data flows associated with applications hosted by the service platform are disclosed. Some of the techniques may include allocating a first amount of bandwidth of a physical underlay of a network for data flows associated with an application. The techniques may also include receiving, from a scalable application service hosting the application, an indication of an amount of computing resources of the scalable application service that are allocated to host the application. Based at least in part on the indications, a second amount of bandwidth of the physical underlay to allocate for the data flows may be determined. The techniques may also include allocating the second amount of bandwidth of the physical underlay of the network for the data flows associated with the application.

Abstract: Presented herein are techniques for updating detailed maps used to navigate an autonomous vehicle. The techniques include determining that a vehicle has come within a predetermined range of a road side unit, establishing a communication link with the vehicle, receiving, from the vehicle, data sufficient to identify a vehicle type of the vehicle, based on the vehicle type, selecting a map, stored by the road side unit, for the vehicle, sending a query to a neighbor road side unit seeking data to augment the map, in response to the query, receiving the data to augment the map from the neighbor road side unit, updating the map based on the data to augment the map to obtain an updated map, and sending at least a aspects of the updated map to the vehicle.

Abstract: Techniques for tracking compute capacity of a scalable application service platform to perform dynamic bandwidth allocation for data flows associated with applications hosted by the service platform are disclosed. Some of the techniques may include allocating a first amount of bandwidth of a physical underlay of a network for data flows associated with an application. The techniques may also include receiving, from a scalable application service hosting the application, an indication of an amount of computing resources of the scalable application service that are allocated to host the application. Based at least in part on the indications, a second amount of bandwidth of the physical underlay to allocate for the data flows may be determined. The techniques may also include allocating the second amount of bandwidth of the physical underlay of the network for the data flows associated with the application.

Abstract: Presented herein are techniques for training a central/global machine learning model in a distributed machine learning system. In the data sampling techniques, a subset of the data obtained at the local sites is intelligently selected for transfer to the central site for use in training the central machine learning model. In the model merging techniques, distributed local training occurs in each local site and copies of the local machine learning models are sent to the central site for aggregation of learning by merging of the models. As a result, in accordance with the examples presented herein, a central machine learning model can be trained based on various representations/transformations of data seen at the local machine learning models, including sampled selections of data-label pairs, intermediate representation of training errors, or synthetic data-label pairs generated by models trained at various local sites.

Abstract: Presented herein are techniques for updating detailed maps used to navigate an autonomous vehicle. The techniques include determining that a vehicle has come within a predetermined range of a road side unit, establishing a communication link with the vehicle, receiving, from the vehicle, data sufficient to identify a vehicle type of the vehicle, based on the vehicle type, selecting a map, stored by the road side unit, for the vehicle, sending a query to a neighbor road side unit seeking data to augment the map, in response to the query, receiving the data to augment the map from the neighbor road side unit, updating the map based on the data to augment the map to obtain an updated map, and sending at least a aspects of the updated map to the vehicle.

Abstract: Presented herein are techniques for automatically generating object segmentation training data. In particular, a segmentation data generation system is configured to obtain training images derived from a scene captured by one or more image capture devices. Each training image is a still image that includes a foreground object and a background. The segmentation data generation system automatically generates a mask of the training image to delineate the object from the background and, based on the mask automatically generates a masked image. The masked image includes only the object present in the training image. The segmentation data generation system composites the masked image with an image of an environmental scene to generate a composite image that includes the masked image and the environmental scene.

Abstract: Presented herein are techniques for updating detailed maps used to navigate an autonomous vehicle. The techniques include determining that a vehicle has come within a predetermined range of a road side unit, establishing a communication link with the vehicle, receiving, from the vehicle, data sufficient to identify a vehicle type of the vehicle, based on the vehicle type, selecting a map, stored by the road side unit, for the vehicle, sending a query to a neighbor road side unit seeking data to augment the map, in response to the query, receiving the data to augment the map from the neighbor road side unit, updating the map based on the data to augment the map to obtain an updated map, and sending at least a aspects of the updated map to the vehicle.

Abstract: A method for encoding a sequence of packets includes receiving the sequence of packets, generating a parity packet for a first group of packets within the sequence of packets, and transmitting the first group of packets and the parity packet. The parity packet is generated by performing an exclusive OR (XOR) operation over a plurality of packets in the first group of packets and at least one packet in a second group of packets. The second group is separated from the first group in the sequence by one or more packets.

Abstract: An endpoint device receives a sequence of audio frames. The endpoint device determines for each audio frame a respective importance level among possible importance levels ranging from a low importance level to a high importance level based on content in the audio frame indicative of the respective importance level. The endpoint device associates each audio frame with the respective importance level, to produce different subsets of audio frames associated with respective ones of different importance levels. The endpoint device, for each subset of audio frames, applies forward error correction to a fraction of audio frames in the subset of audio frames, wherein the fraction increases as the importance level of the audio frames in the subset increases, and does not apply forward error correction to remaining audio frames in the subset.

Abstract: Presented herein are techniques for automatically generating object segmentation training data. In particular, a segmentation data generation system is configured to obtain training images derived from a scene captured by one or more image capture devices. Each training image is a still image that includes a foreground object and a background. The segmentation data generation system automatically generates a mask of the training image to delineate the object from the background and, based on the mask automatically generates a masked image. The masked image includes only the object present in the training image. The segmentation data generation system composites the masked image with an image of an environmental scene to generate a composite image that includes the masked image and the environmental scene.

Abstract: Presented herein are techniques for assignment of an identity to a group of captured images. A plurality of captured images that each include an image of at least one person are obtained. For each of the plurality of captured images, relational metrics indicating a relationship between the image of the person in a respective captured image and the images of the persons in each of the remaining plurality of captured images is calculated. Based on the relational metrics, a clustering process is performed to generate one or more clusters from the plurality of captured images. Each of the one or more clusters are associated with an identity of an identity database. The one or more clusters may each be associated with an existing identity of the identity database or an additional identity that is not yet present in the identity database.

Abstract: Presented herein are techniques for automatically generating object segmentation training data. In particular, a segmentation data generation system is configured to obtain training images derived from a scene captured by one or more image capture devices. Each training image is a still image that includes a foreground object and a background. The segmentation data generation system automatically generates a mask of the training image to delineate the object from the background and, based on the mask automatically generates a masked image. The masked image includes only the object present in the training image. The segmentation data generation system composites the masked image with an image of an environmental scene to generate a composite image that includes the masked image and the environmental scene.

Abstract: An endpoint device receives a sequence of audio frames. The endpoint device determines for each audio frame a respective importance level among possible importance levels ranging from a low importance level to a high importance level based on content in the audio frame indicative of the respective importance level. The endpoint device associates each audio frame with the respective importance level, to produce different subsets of audio frames associated with respective ones of different importance levels. The endpoint device, for each subset of audio frames, applies forward error correction to a fraction of audio frames in the subset of audio frames, wherein the fraction increases as the importance level of the audio frames in the subset increases, and does not apply forward error correction to remaining audio frames in the subset.

Abstract: Presented herein are techniques for updating detailed maps used to navigate an autonomous vehicle. The techniques include determining that a vehicle has come within a predetermined range of a road side unit, establishing a communication link with the vehicle, receiving, from the vehicle, data sufficient to identify a vehicle type of the vehicle, based on the vehicle type, selecting a map, stored by the road side unit, for the vehicle, sending a query to a neighbor road side unit seeking data to augment the map, in response to the query, receiving the data to augment the map from the neighbor road side unit, updating the map based on the data to augment the map to obtain an updated map, and sending at least a aspects of the updated map to the vehicle.

Abstract: Presented herein are techniques for training a central/global machine learning model in a distributed machine learning system. In the data sampling techniques, a subset of the data obtained at the local sites is intelligently selected for transfer to the central site for use in training the central machine learning model. In the model merging techniques, distributed local training occurs in each local site and copies of the local machine learning models are sent to the central site for aggregation of learning by merging of the models. As a result, in accordance with the examples presented herein, a central machine learning model can be trained based on various representations/transformations of data seen at the local machine learning models, including sampled selections of data-label pairs, intermediate representation of training errors, or synthetic data-label pairs generated by models trained at various local sites.

Abstract: Presented herein are techniques for assignment of an identity to a group of captured images. A plurality of captured images that each include an image of at least one person are obtained. For each of the plurality of captured images, relational metrics indicating a relationship between the image of the person in a respective captured image and the images of the persons in each of the remaining plurality of captured images is calculated. Based on the relational metrics, a clustering process is performed to generate one or more clusters from the plurality of captured images. Each of the one or more clusters are associated with an identity of an identity database. The one or more clusters may each be associated with an existing identity of the identity database or an additional identity that is not yet present in the identity database.

Abstract: Presented herein are techniques for automatically generating object segmentation training data. In particular, a segmentation data generation system is configured to obtain training images derived from a scene captured by one or more image capture devices. Each training image is a still image that includes a foreground object and a background. The segmentation data generation system automatically generates a mask of the training image to delineate the object from the background and, based on the mask automatically generates a masked image. The masked image includes only the object present in the training image. The segmentation data generation system composites the masked image with an image of an environmental scene to generate a composite image that includes the masked image and the environmental scene.

Abstract: An example partially transmissive display is provided to output data to a user. The partially transmissive display can include a capture device to receive captured data transmitted through the partially transmissive display, and data generated on and/or reflected from the partially transmissive display. A cross-talk reduction component may be included to receive the captured data and a remote signal comprising displayed content on the partially transmissive display. The cross-talk component may utilize the remote signal and the captured data to reduce cross-talk and output a corrected signal to one or more remote participants.