Abstract: A data compression apparatus is described which has an encoder configured to receive an input data item and to compress the data item into an encoding comprising a plurality of numerical values. The numerical values are grouped at least according to whether they relate to content of the input data item or style of the input data item. The encoder has been trained using a plurality of groups of training data items grouped according to the content and where training data items within individual ones of the groups vary with respect to the style. The encoder has been trained using a training objective which takes into account the groups.

Abstract: A data compression apparatus is described which has an encoder configured to receive an input data item and to compress the data item into an encoding comprising a plurality of numerical values. The numerical values are grouped at least according to whether they relate to content of the input data item or style of the input data item. The encoder has been trained using a plurality of groups of training data items grouped according to the content and where training data items within individual ones of the groups vary with respect to the style. The encoder has been trained using a training objective which takes into account the groups.

Abstract: In one embodiment, network information associated with a plurality of nodes in a network is received at a device in a network. From the plurality of nodes, a node is selected based on a determination that the selected node is an outlier among the plurality of nodes according to the received network information. Then, a probe is sent to the selected node, and in response to the probe, a performance metric is received from the selected node at the device.

Abstract: A data compression apparatus is described which has an encoder configured to receive an input data item and to compress the data item into an encoding comprising a plurality of numerical values. The numerical values are grouped at least according to whether they relate to content of the input data item or style of the input data item. The encoder has been trained using a plurality of groups of training data items grouped according to the content and where training data items within individual ones of the groups vary with respect to the style. The encoder has been trained using a training objective which takes into account the groups.

Abstract: A data compression apparatus is described which has an encoder configured to receive an input data item and to compress the data item into an encoding comprising a plurality of numerical values. The numerical values are grouped at least according to whether they relate to content of the input data item or style of the input data item. The encoder has been trained using a plurality of groups of training data items grouped according to the content and where training data items within individual ones of the groups vary with respect to the style. The encoder has been trained using a training objective which takes into account the groups.

Abstract: In one embodiment, network information associated with a plurality of nodes in a network is received at a device in a network. From the plurality of nodes, a node is selected based on a determination that the selected node is an outlier among the plurality of nodes according to the received network information. Then, a probe is sent to the selected node, and in response to the probe, a performance metric is received from the selected node at the device.

Abstract: In one embodiment, a central device receives a routing strategy instruction that specifies a predictability threshold for communication delays in the network. The device estimates communication delays for a plurality of paths in the network and determines predictability measurements for the estimated delays. The device also selects, from among the plurality of paths, a particular path that has a predictability measurement that satisfies the predictability threshold and has a minimal estimated delay. The central device further installs the particular path at one or more other devices in the network.

Abstract: In one embodiment, network information associated with a plurality of nodes in a network is received at a device in a network. From the plurality of nodes, a node is selected based on a determination that the selected node is an outlier among the plurality of nodes according to the received network information. Then, a probe is sent to the selected node, and in response to the probe, a performance metric is received from the selected node at the device.

Abstract: In one embodiment, a machine learning model for predicting one or more metrics is run in a network which includes a centralized controller device interconnected with a plurality of edge devices. A batch version of the machine learning model that operates in batch mode is hosted at the centralized controller device. Then, an incremental version of the machine learning model that operates in incremental mode is pushed to an edge device of the plurality of edge devices, such that the incremental version of the machine learning model is hosted at the edge device. As a result, the batch version and the incremental version of the machine learning model run in parallel with one another.

Abstract: In one embodiment, a method is disclosed in which a device receives delay information for a communication segment in a network. The device determines a predictability measurement for delays along the segment using the received delay information. The predictability measurement is advertised to one or more devices in the network and used as a routing constraint to select a routing path in the network.

Abstract: In one embodiment, network information associated with a plurality of nodes in a network is received at a device in a network. From the plurality of nodes, a node is selected based on a determination that the selected node is an outlier among the plurality of nodes according to the received network information. Then, a probe is sent to the selected node, and in response to the probe, a performance metric is received from the selected node at the device.

Abstract: In one embodiment, a machine learning model for predicting one or more metrics is run in a network which includes a centralized controller device interconnected with a plurality of edge devices. A batch version of the machine learning model that operates in batch mode is hosted at the centralized controller device. Then, an incremental version of the machine learning model that operates in incremental mode is pushed to an edge device of the plurality of edge devices, such that the incremental version of the machine learning model is hosted at the edge device. As a result, the batch version and the incremental version of the machine learning model run in parallel with one another.

Abstract: In one embodiment, a central device receives a routing strategy instruction that specifies a predictability threshold for communication delays in the network. The device estimates communication delays for a plurality of paths in the network and determines predictability measurements for the estimated delays. The device also selects, from among the plurality of paths, a particular path that has a predictability measurement that satisfies the predictability threshold and has a minimal estimated delay. The central device further installs the particular path at one or more other devices in the network.

Abstract: In one embodiment, a method is disclosed in which a device receives delay information for a communication segment in a network. The device determines a predictability measurement for delays along the segment using the received delay information. The predictability measurement is advertised to one or more devices in the network and used as a routing constraint to select a routing path in the network.