Abstract: System and method for layer-wise training of deep neural networks (DNNs) are disclosed. In an embodiment, multiple labelled images are received at a layer of multiple layers of a DNN. Further, the labelled images are pre-processed. The pre-processed images are then transformed based on a predetermined weight matrix to obtain feature representation of the pre-processed images at the layer, the feature representation comprise feature vectors and associated labels. Furthermore, kernel similarity between the feature vectors is determined based on a predefined kernel function. Moreover, a Gaussian kernel matrix is determined based on the kernel similarity. In addition, an error function is computed based on the predetermined weight matrix and the Gaussian kernel matrix. Also, a weight matrix associated with the layer is computed based on the error function and predetermined weight matrix, thereby training the layer of the multiple layers.

Abstract: In a decoupled messaging system, a producer device sends produced messages to a message fabric. The message fabric either stores the received message in designated queues, from which consumer devices may retrieve the messages, or broadcasts the messages to consumer devices that subscribe to topics assigned to the messages. A message consumption orchestrator service gathers capabilities information for the message producer, the message fabric, and the message consumer, and configures resources allocated for the message consumer to more fully optimize message throughput. Further, the message consumption orchestrator identifies patterns in changes of the rate of produced messages, i.e., through analysis of historical data and/or received information. The message consumption orchestrator configures resources for the message consumer to account for predicted need based on the identified patterns.

Abstract: An image display method is provided. The image display method includes displaying a camera preview screen that provides an image of a subject in real time, capturing one or more images by capturing the camera preview screen, and displaying a photo reel on the camera preview screen, the photo reel being generated by arranging thumbnails of the captured one or more images in a row.

Abstract: The present application provides a method and system for unsupervised word image clustering, comprises capturing one or more image wherein the one or more image comprises at least one word images. Extracting at least one feature vector using an untrained convolution neural network architecture, wherein the convolution filters are initialized by random filter based deep learning techniques using Gaussian random variable with zero mean and unit standard deviation, and wherein the convolution filters are constrained to sum to zero. The extracted feature vectors are used for clustering, wherein clustering is performed in two stages. First stage includes clustering word images which are similar using a graph connected component. Second stage clustering includes clustering a remaining word images which are not clustered during the first stage by evaluating the remaining images against the clusters formed during the first stage and assigning them to clusters based on the evaluation.

Abstract: System and method for layer-wise training of deep neural networks (DNNs) are disclosed. In an embodiment, multiple labelled images are received at a layer of multiple layers of a DNN. Further, the labelled images are pre-processed. The pre-processed images are then transformed based on a predetermined weight matrix to obtain feature representation of the pre-processed images at the layer, the feature representation comprise feature vectors and associated labels. Furthermore, kernel similarity between the feature vectors is determined based on a predefined kernel function. Moreover, a Gaussian kernel matrix is determined based on the kernel similarity. In addition, an error function is computed based on the predetermined weight matrix and the Gaussian kernel matrix. Also, a weight matrix associated with the layer is computed based on the error function and predetermined weight matrix, thereby training the layer of the multiple layers.

Abstract: An electronic device includes a touch interface and a processor configured to identify a current touch displacement while a touch gesture of a user is received through the touch interface, to identify an action corresponding to a current state of the electronic device and the identified current touch displacement, and to perform a task according to the identified action.

Abstract: A system, method, and computer-readable medium, including obtaining a plurality of different application program interfaces (APIs), one for each of a plurality of different technology stacks, to be used in a test flow of a product constructed of the plurality of different technology stacks; generating a model of a test flow using each of the plurality of different APIs; defining input and output parameters for each of the plurality of different APIs in the model; automatically generating executable instructions to execute the modeled test flow; and storing a record of the executable instructions to execute the modeled test flow.

Abstract: A system, method, and computer-readable medium, including obtaining a plurality of different application program interfaces (APIs), one for each of a plurality of different technology stacks, to be used in a test flow of a product constructed of the plurality of different technology stacks; generating a model of a test flow using each of the plurality of different APIs; defining input and output parameters for each of the plurality of different APIs in the model; automatically generating executable instructions to execute the modeled test flow; and storing a record of the executable instructions to execute the modeled test flow.

Abstract: The disclosure generally relates to transcription of spoken words, and more particularly to a system and method for transcription of spoken words using multilingual mismatched words. The process comprises collection of multi-scripted noisy transcriptions of the spoken word obtained from workers of the multilingual mismatched crowd. The collected words are mapped to a phoneme sequence in the source language using script specific graphemes to phoneme model. Further, it builds a multi-scripted transcription script specific, worker specific and a global insertion-deletion-substitution (IDS) channel. Furthermore, the disclosure also determines reputation of workers to allocate the transcription task. Determination of reputation is based on word belief.

Abstract: A solar monitoring system for measuring solar radiation intensity comprising a tracking unit having two-axis movement comprising, an image capturing head mounted with first and second irradiation measuring units, and a controller. The first irradiation measuring unit comprises a direct normal irradiance (DNI) sensor and the second irradiation measuring unit includes a diffuse horizontal irradiance (DHI) sensor and a global horizontal irradiance (GHI) sensor. The controller receives inputs from the sensors or a software program configured to control orientation of the image capturing head so that the DNI sensor is always exposed to the sun, and the shading disc is always directly between the DHI sensor and the sun.

Abstract: Visual representations of gadgets, each of which is configured to perform a different function, are displayed by an electronic device. The gadgets are arranged in a sequence to build a workflow. The gadgets are executed in the sequence to execute the workflow on a project.

Abstract: A mechanism is provided for controlling a process on a computing node. In one implementation, a determination is made as to whether the participation status of a computing node has changed. The participation status may, for example, be active, in which case the computing node is participating in the processing of requests from a request handler, or it may be passive, in which case the computing node is not participating in the processing of requests from the request handler. In response to a determination that the participation status of the computing node has changed from a current status to a new status, one or more processes on the computing node may be caused to begin or halt processing, in accordance with the new status. For example, if the current status is passive and the new status is active, a process on the computing node may be caused to begin processing.

Abstract: Techniques are described for identifying unhealthy nodes in a multi-node system. One or more parameters of each node is monitored, then compared with the values for the same parameter running on other nodes in the multi-node system. Based on the comparison, a determination is made whether a node is healthy. If the multi-node system comprises one or more nodes with differing capabilities, an adjustment is performed to account for the differing capabilities of each respective node. Further provided are methods of taking remedial action upon a determination that a node is unhealthy. A tuner is used to modify values of health parameters until the node is performing similarly to its peers.

Abstract: The present application provides a method and system for unsupervised word image clustering, comprises capturing one or more image wherein the one or more image comprises at least one word images. Extracting at least one feature vector using an untrained convolution neural network architecture, wherein the convolution filters are initialized by random filter based deep learning techniques using Gaussian random variable with zero mean and unit standard deviation, and wherein the convolution filters are constrained to sum to zero. The extracted feature vectors are used for clustering, wherein clustering is performed in two stages. First stage includes clustering word images which are similar using a graph connected component. Second stage clustering includes clustering a remaining word images which are not clustered during the first stage by evaluating the remaining images against the clusters formed during the first stage and assigning them to clusters based on the evaluation.

Abstract: A processor-implemented method and system for detecting objects in a media file using a block based histogram of oriented gradients methodology is described. At least one image representing at least one frame of a video sequence comprising one or more objects of at least one desired object type is received. A region of interest contained in a block of frame is considered. An array containing magnitude and angle values is calculated, normalized for each pixel in the block in the internal memory. The normalized array is aggregated, ported and stored in an external memory. A value or block of values of the aggregated array in the external memory along with the specific feature vector is transferred to the internal memory. Based on a comparison of each value in the aggregated array and the specific feature vector, a desired object type is detected.

Abstract: Optimizing task allocation requires taking into account cognitive load on workers and their response time to allocated tasks. The present disclosure provides for allocation of task by receiving data pertaining to current activity of workers; receiving data pertaining to at least one task to be allocated and determining activity-task pairs based on an activity feature vector corresponding to at least one human body part used during the current activity and a task feature vector corresponding to at least one human body part required for the at least one task to be performed by the workers. Cognitive load on the workers is then estimated for the determined activity-task pairs. An optimum activity-task pair based on the estimated cognitive load is determined and at least one task is allocated to the workers based on the determined optimum activity-task pair.

Abstract: A processor-implemented method and system for detecting objects in a media file using a block based histogram of oriented gradients methodology is described. At least one image representing at least one frame of a video sequence comprising one or more objects of at least one desired object type is received. A region of interest contained in a block of frame is considered. An array containing magnitude and angle values is calculated, normalized for each pixel in the block in the internal memory. The normalized array is aggregated, ported and stored in an external memory. A value or block of values of the aggregated array in the external memory along with the specific feature vector is transferred to the internal memory. Based on a comparison of each value in the aggregated array and the specific feature vector, a desired object type is detected.

Abstract: In a decoupled messaging system, a producer device sends produced messages to a message fabric. The message fabric either stores the received message in designated queues, from which consumer devices may retrieve the messages, or broadcasts the messages to consumer devices that subscribe to topics assigned to the messages. A message consumption orchestrator service gathers capabilities information for the message producer, the message fabric, and the message consumer, and configures resources allocated for the message consumer to more fully optimize message throughput. Further, the message consumption orchestrator identifies patterns in changes of the rate of produced messages, i.e., through analysis of historical data and/or received information. The message consumption orchestrator configures resources for the message consumer to account for predicted need based on the identified patterns.

Abstract: In a decoupled messaging system, a producer device sends produced messages to a message fabric. The message fabric either stores the received message in designated queues, from which consumer devices may retrieve the messages, or broadcasts the messages to consumer devices that subscribe to topics assigned to the messages. A message consumption orchestrator service gathers capabilities information for the message producer, the message fabric, and the message consumer, and configures resources allocated for the message consumer to more fully optimize message throughput. Further, the message consumption orchestrator identifies patterns in changes of the rate of produced messages, i.e., through analysis of historical data and/or received information. The message consumption orchestrator configures resources for the message consumer to account for predicted need based on the identified patterns.

Abstract: The present discussion relates to generating power generation forecasts both on-site and remote to a wind farm, or other intermittent power generation asset, so as to increase the reliability of providing a forecast to interested parties, such as regulatory authorities. Forecasts may be separately generated at both the on-site and remote locations and, if both are available, one is selected for transmission to interested parties, such as regulatory authorities. If, due to circumstances, one forecast is unavailable, the other forecast may be used in its place locally and remotely, communications permitting.