Abstract: In one embodiment, an apparatus comprises a communication interface and a processor. The communication interface is to communicate with a plurality of devices. The processor is to: receive compressed data from a first device, wherein the compressed data is associated with visual data captured by sensor(s); perform a current stage of processing on the compressed data using a current CNN, wherein the current stage of processing corresponds to one of a plurality of processing stages associated with the visual data, and wherein the current CNN corresponds to one of a plurality of CNNs associated with the plurality of processing stages; obtain an output associated with the current stage of processing; determine, based on the output, whether processing associated with the visual data is complete; if the processing is complete, output a result associated with the visual data; if the processing is incomplete, transmit the compressed data to a second device.

Abstract: Video coding techniques including differential bit rate or quality coding of one or more regions of interest and one or more non-regions of interest based on information including one or more of coordinates of the one or more regions of interest, a target complexity, residual encoder bit data, a requested quality, a difference between the current video data frame and a reconstructed video data frame, a target quality, a requested bit rate, frame target bit allocation and an as encoded bit rate.

Abstract: A method and apparatus for characteristic-based video processing include: in response to receiving a region of a picture of a video sequence, determining a characteristic in the region, the region being independent of other regions of the picture for video coding; determining a class associated with the region based on the characteristic, the class being selected from a plurality of classes; and encoding the region using a parameter set associated with the class, the parameter set being selected from a plurality of parameter sets for video coding at different quality levels.

Abstract: Data is received describing a local model of a first device generated by the first device based on sensor readings at the first device and a global model is updated that is hosted remote from the first device based on the local model and modeling devices in a plurality of different asset taxonomies. A particular operating state affecting one or more of a set of devices deployed in a particular machine-to-machine network is detected and the particular machine-to-machine network is automatically reconfigured based on the global model.

Abstract: In one embodiment, an apparatus comprises a memory and a processor. The memory is to store visual data associated with a visual representation captured by one or more sensors. The processor is to: obtain the visual data associated with the visual representation captured by the one or more sensors, wherein the visual data comprises uncompressed visual data or compressed visual data; process the visual data using a convolutional neural network (CNN), wherein the CNN comprises a plurality of layers, wherein the plurality of layers comprises a plurality of filters, and wherein the plurality of filters comprises one or more pixel-domain filters to perform processing associated with uncompressed data and one or more compressed-domain filters to perform processing associated with compressed data; and classify the visual data based on an output of the CNN.

Abstract: Methods and systems are provided for implementing static channel filtering operations upon image datasets transformed to frequency domain representations, including decoding images of an image dataset to generate a frequency domain representation of the image dataset; discarding coefficient values of one or more particular frequency channels of each image of the image dataset in a frequency domain representation; and transporting the image dataset in a frequency domain representation to one or more special-purpose processor(s). Methods and systems of the present disclosure may enable a filtered image dataset to be input to a second layer of a learning model, bypassing a first layer, or may enable a learning model to be designed with a reduced-size first layer. This may achieve benefits such as reducing computational overhead and time of machine learning training and inference computations, reducing volume of image data input into the learning model, and reducing convergence time.

Abstract: Methods and systems are provided for implementing preprocessing operations and augmentation operations upon image datasets transformed to frequency domain representations, including decoding images of an image dataset to generate a frequency domain representation of the image dataset; performing a resizing operation based on resizing factors on the image dataset in a frequency domain representation; performing a reshaping operation based on reshaping factors on the image dataset in a frequency domain representation; and performing a cropping operation on the image dataset in a frequency domain representation. The methods and systems may further include performing an augmentation operation on the image dataset in a frequency domain representation. Methods and systems of the present disclosure may free learning models from computational overhead caused by transforming image datasets into frequency domain representations.

Abstract: In one embodiment, an apparatus comprises a storage device and a processor. The storage device may store a plurality of compressed images comprising one or more compressed master images and one or more compressed slave images.

Abstract: Multiple devices are detected in an environment and a user input is received to define a relationship between two or more devices in the plurality of devices. A system can determine that a first of the two or more devices includes a sensor resource and a second of the two or more devices includes an actuator resource. Data is identified describing outputs of the first device corresponding to the sensor resource and inputs of the second device corresponding to the actuator resource. A model is generated modeling interoperation of the sensor resource and actuator resource based at least in part on the data.

Abstract: Image data is accessed. The image data includes frequency domain components. A subset of the frequency domain components is selected based on the relative importance of the frequency domain components. Only the subset of the frequency domain components is provided to an accelerator that executes a neural network to perform an artificial intelligence task using the subset of frequency domain components.

Abstract: A system for determining the importance of encoded image components for artificial intelligence tasks includes an image capture or storage unit, a processor and a communication interface. The processor can receive components of transformed domain image data from the one or more image capture or storage units across the communication interface. The processor can be configured to determine the relative importance of the components of the transformed domain image data for an artificial intelligence task.

Abstract: Discrete cosine transformation (DCT) information can be estimated from adjacent blocks of the same frame. DCT information can be estimated from different frames. Motion vectors can be used to track the position of objects in some frames of the video. For example, a stream of encoded frames is received; the encoded frames are entropy decoded and dequantized to produce DCT information for blocks of the frames; and DCT information for a block in a frame is determined using the DCT information produced from the entropy decoding and dequantizing for a different block.

Abstract: A system for processing encoded image components for artificial intelligence tasks. The system can include one or more compute units, one or more controllers and memory. The one or more controllers can include one or more micro-op schedulers and one or more channel switches. The one or more compute units can be configured to process components of the transformed domain image data according to one or more micro-operations for an artificial intelligence task. The one or more channel switches can be configured to selectively control the transfer of the components of transformed domain image data to the one or more compute units based on one or more gating flags. The one or more channel switches can also be configured to selectively control generation of the one or more micro-operations by the one or more micro-op schedulers based on the one or more gating flags.

Abstract: Availability of computing resources is detected on a particular device in a network and a runtime core is caused to be loaded on the particular device based on the availability. The runtime core is configured to support hot-plugging of code embodying any one of a plurality of job and first code comprising a placeholder job is caused to be run on the runtime core to reserve at least a portion of the computing resources of the particular device. A particular one of the plurality of jobs to be run on the particular device is identified and the first code is replaced with second code corresponding to the particular job to replace the placeholder job on the runtime core.

Abstract: A user input including an identification of a set of job abstractions is received, where each job abstraction in the set of job abstractions includes a respective one of a plurality of defined job abstractions and each of the plurality of defined job abstractions are mapped to two or more asset capability abstractions in a plurality of defined asset capability abstractions. The user input is processed to generate program data, based on the set of job abstractions. The resulting program data is executable by a processor device to: identify a set of asset capability abstractions in the plurality of asset capability abstractions corresponding to the set of job abstractions; determine that a set of devices in an environment possess capabilities corresponding to the set of asset capability abstractions; and launch a system including the set of devices to implement jobs corresponding to the set of job abstractions.

Abstract: The systems and methods are configured to efficiently and effectively determine or find an estimated optimal encoding parameter set. In one embodiment, a video encoding parameter set estimation method comprises: performing an offline encoding parameter set characteristic prediction process that determines an estimate of a candidate encoding parameter set characteristic; and performing an encoding parameter set search process that identifies a predicted or estimated optimal video encoding parameter set. The encoding parameter set search process can include applying a constraint to the candidate encoding parameter set characteristic; and determining if candidate encoding parameter set meets an objective, wherein the determining is performed if the constraint is satisfied. The candidate encoding parameter set characteristic can be an estimated encoding time of the candidate encoding parameter set. The objective can be the best video quality out of a plurality of candidate encoding parameter sets.

Abstract: A plurality of devices are detected within range of a particular device, and capabilities of each of the plurality of devices are determined, as well as a respective taxonomy to be associated with each device based on the device's capabilities. A set of asset abstractions are identified, referenced by a particular software application configured to manage a particular machine-to-machine network. Each asset abstraction can correspond to a respective one of the set of taxonomies, and the particular machine-to-machine network can be built from an instance of each one of the set of asset abstractions. A subset of the plurality of devices can be selected for deployment in the particular machine-to-machine network based on the associated taxonomies, the subset of devices representing an instance of each one of the set of asset abstractions. The subset of devices are then automatically deployed to implement the instance of the particular machine-to-machine network.

Abstract: A mechanism is described for facilitating portable, reusable, and shareable Internet of Things-based services and resources according to one embodiment. A method of embodiments, as described herein, includes wherein receiving a recipe request for selecting a recipe relating to Internet of Things (IoT) services, where selecting includes at least one of searching the recipe and modifying the recipe. The method may further include selecting the recipe, where the recipe includes a set of resource requirements and business logic. The method may further include modifying the set of resource requirements, where the modified set of resource requirements is associated with the business logic to modify the recipe, and deploying the modified recipe at one or more computing devices, where the modified recipe to facilitate management for one or more IoT devices at one or more locations.

Abstract: Systems, apparatuses and methods may identify a capability abstraction in a request to configure a first Internet of Things (IOT) application in a physical environment including a plurality of IOT devices and select a resource abstraction from a plurality of resource abstractions based on the capability abstraction. The selected resource abstraction may correspond to a first IOT device in the plurality of IOT devices. Additionally, the first IOT application may be bound with the first IOT device. In one example, first data originating from the first IOT device is received, a first runtime abstraction is selected from a plurality of runtime abstractions, wherein the first runtime abstraction corresponds to the first IOT application, and the first data is sent to the first IOT application via the first runtime abstraction.

Abstract: In one embodiment, an apparatus comprises a processor to: identify a workload comprising a plurality of tasks; generate a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks; identify a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with a plurality of processing devices; identify a privacy policy associated with the workload; identify privacy level information associated with the plurality of processing devices; identify a privacy constraint based on the privacy policy and the privacy level information; and determine a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph.