Abstract: Devices, systems, and methods are provided for generating and selecting video clips for inclusion in video sequences based on still frame images. A method may include encoding first embeddings for a first video including first images of an item at a first scene, the first embeddings indicative of features of the first scene; encoding second embeddings for a second video including second images of the item at a second scene, the second embeddings indicative of features of the second scene; encoding third embeddings for the first video, the third embeddings indicative of features of a first type of camera shot used for the first images; encoding fourth embeddings for the second video, the fourth embeddings indicative of features of a second type of camera shot used for the second images; and generating, based on the first, second, third, and fourth embeddings, a video sequence for the item.

Abstract: Technologies are provided for automated generation of directed content campaigns. The generated campaign can be optimized for performance. In some embodiments, a group of variations of attributes that define a directed content campaign can be generated and allocated traffic weights for respective impressions of the directed content campaign in a media outlet channel. The traffic weights can then be iteratively updated until a termination criterion is satisfied. At each iteration, the traffic weight can be updated by applying a machine-learning model to current performance metric values of respective impressions corresponding to the traffic weights. After termination of the updates to the traffic weights, a particular set of variations having traffic weights exceeding a threshold can be selected as directed content campaign having satisfactory performance. Those variations can be supplied to a requestor device for subsequent utilization.

Abstract: Techniques are generally described for context aware text-to-image synthesis. First text data comprising a description of an object may be received. A recurrent neural network may determine a first semantic representation data representing the first text data. A generator trained using a first generative adversarial network (GAN) may determine first image data representing the object using the first semantic representation. An encoder of a second GAN may generate a first feature representation of the first image data. The first feature representation may be combined with a projection of the first semantic representation data. A decoder of the second GAN may generate second image data representing the first text data.

Abstract: Methods and arrangements for identifying burden comprehension in multimedia content. A contemplated method includes: accepting multimedia input; detecting components of the multimedia input; determining a comprehension burden score of each of the detected components; and thereupon calculating a total comprehension burden score for the multimedia input. Other variants and embodiments are broadly contemplated herein.

Abstract: Methods and arrangements for generating a learning graph. A contemplated method includes: utilizing at least one processor to execute instructions to perform the steps of: receiving a proficiency input relating to a student; receiving a target knowledge node, wherein the target knowledge node represents at least one skill the student does not currently possess; determining at least one skill requirement of the at least one skill; identifying at least one path between the proficiency input and the target knowledge node based upon the at least one determined skill requirement; calculating a gap between the proficiency input and the target knowledge node at the at least one identified path; and recommending at least one learning content module based upon the calculated gap.

Abstract: The disclosed embodiments illustrate methods of generation of a table of content by processing multimedia content. The method includes identifying a set of key-phrases from the multimedia content based on one or more external data sources. The method further includes determining one or more segments of the multimedia content, based on the identified set of key-phrases, wherein a segment of the determined one or more segments comprises a subset of key-phrases from the set of key-phrases. The method further includes selecting at least a key-phrase from the subset of key-phrases of each of the corresponding one or more segments. The method further includes generating the table of content based on the selected key-phrase from each of the one or more segments, wherein the selected key-phrase from each of the one or more segments in the generated table of content is utilized to navigate through the multimedia content.

Abstract: Features are extracted from visual and audio modalities of a video to infer the location of figures/tables/equations/graphs/flow-charts determined as video anchor points which are highlighted on the video timeline to enable quick navigation and provide a quick summary of the video. A voice-based mechanism navigates to a point-of-interest in the video. In case of bandwidth-constrained settings, videos are often played at a very low resolution (quality), and often users need to increase video resolution manually to understand content presented in the figures. Using the automatic identification of these aforementioned anchored points, the resolution can be changed dynamically during streaming a video, which will provide a better viewing experience.

Abstract: The disclosed embodiments illustrate method and system of processing multimedia content to generate a text transcript. The method includes segmenting each of a set of text frames to determine spatial regions. The method further includes extracting one or more keywords from each of the determined spatial regions. The method further includes determining the first set of keywords from the extracted one or more keywords based on filtering of one or more off-topic keywords from the extracted one or more keywords. The method further includes extracting a second set of keywords based on the determined first set of keywords. The method further includes generating a graph between each of a first set of keywords and one or more of a second set of keywords. The method further includes dynamically generating the text transcript of audio content in the multimedia content based on the generated graph.

Abstract: The disclosed embodiments illustrate methods and systems for querying multiple healthcare-related knowledge graphs. The method includes retrieving a set of healthcare-related response sub-graphs from a plurality of healthcare-related knowledge graphs based on a keyword-based query. The method further includes generating a first set of healthcare-related ranked sub-graphs corresponding to the plurality of healthcare-related knowledge graphs. The method further includes generating a set of healthcare-related connected sub-graphs, based on at least healthcare-related ranked sub-graphs in the plurality of healthcare-related knowledge graphs. The method further includes generating a second set of healthcare-related ranked sub-graphs based on at least a ranking of healthcare-related connected sub-graphs in the set of healthcare-related connected sub-graphs.

Abstract: The disclosed embodiments illustrate method and system of processing multimedia content to generate a text transcript. The method includes segmenting each of a set of text frames to determine spatial regions. The method further includes extracting one or more keywords from each of the determined spatial regions. The method further includes determining the first set of keywords from the extracted one or more keywords based on filtering of one or more off-topic keywords from the extracted one or more keywords. The method further includes extracting a second set of keywords based on the determined first set of keywords. The method further includes generating a graph between each of a first set of keywords and one or more of a second set of keywords. The method further includes dynamically generating the text transcript of audio content in the multimedia content based on the generated graph.

Abstract: The disclosed embodiments illustrate method for rendering time-compressed multimedia content on a user-computing device. The method includes determining metadata for one or more frames in multimedia content based on each of one or more time-compression factors and one or more attributes of the multimedia content. Further, the determined metadata comprises a binary value associated with each of the one or more frames of the multimedia content. The method further includes transmitting the multimedia content associated with the determined metadata to the user-computing device, based at least on a time-compression factor in a user request received from the user-computing device. Further, the transmitted multimedia content is rendered on the user-computing device as the time-compressed multimedia content.

Abstract: The disclosed embodiments illustrate methods of generation of a table of content by processing multimedia content. The method includes identifying a set of key-phrases from the multimedia content based on one or more external data sources. The method further includes determining one or more segments of the multimedia content, based on the identified set of key-phrases, wherein a segment of the determined one or more segments comprises a subset of key-phrases from the set of key-phrases. The method further includes selecting at least a key-phrase from the subset of key-phrases of each of the corresponding one or more segments. The method further includes generating the table of content based on the selected key-phrase from each of the one or more segments, wherein the selected key-phrase from each of the one or more segments in the generated table of content is utilized to navigate through the multimedia content.

Abstract: An image display system and method dynamically adjusts a resolution of a streamed image corresponding to determined visual saliency scores of the streamed image. A viewer display, a resolution adaptation engine and a visual saliency score calculation engine are included.

Abstract: Features are extracted from visual and audio modalities of a video to infer the location of figures/tables/equations/graphs/flow-charts determined as video anchor points which are highlighted on the video timeline to enable quick navigation and provide a quick summary of the video. A voice-based mechanism navigates to a point-of-interest in the video. In case of bandwidth-constrained settings, videos are often played at a very low resolution (quality), and often users need to increase video resolution manually to understand content presented in the figures. Using the automatic identification of these aforementioned anchored points, the resolution can be changed dynamically during streaming a video, which will provide a better viewing experience.

Abstract: Systems and methods for identifying entities sharing a temporal pattern using bipartite graphs are described. In one embodiment, a method includes identifying a temporal pattern in a sequence of bipartite graphs for a sequence of records involving two entity types, where records of the sequence of bipartite graphs vary according to time. An embodiment may color code the edges between entity types in the sequence of bipartite graphs according to the at least one temporal pattern identified (e.g., increasing sales between a business representative and a customer). An embodiment may therefore identify a time-based relationship between at least two entities according to the coded edges. Given the identification of entities having a time-based relationships, groups of these entities may be identified and trends may be derived therefrom (e.g., increasing sales for business units of a particular geographic region).

Abstract: An approach is provided for creating a new document. Keywords specifying a subject matter of the new document are received. Metadata of documents is determined to match keyword(s) included in the received keywords and the documents are retrieved. Based on a section being created in the new document, a ranked list of the retrieved documents is generated. A selection of a document included in the ranked list is received. The selected document is added to the new document. The keywords are refined based on the added document. Based on the subject matter and the refined keywords, the new document is completed by repeating the steps of determining the metadata, retrieving the documents, generating the ranked list, receiving the selection, and adding the selected document.

Abstract: Computer software is disclosed for discovering and representing a reporting model of an existing reporting environment. For each report in a plurality of reports, the software searches metadata of the report for descriptive information and dependencies on other reports. The software depicts, in a graphical representation, each report and relationships between the reports.

Abstract: A computer receives one or more files having configuration information that includes data that defines a plurality of stages of an extract, transform, and load (ETL) job, wherein the plurality of stages comprise a read stage that is preceded by a write stage, and wherein the read stage reads data from a source location, and wherein the data that is read or a modified version of the data that is read is being written by the write stage that writes data to the source location. The computer replaces the read stage with a decompressor stage. The computer replaces the write stage with a compressor stage. The computer executes the decompressor stage and compressor stage on a field-programmable gate array that is programmatically customized with data compression and data decompression functionality to enhance the performance of the ETL job.

Abstract: In the context of data administration in enterprises, an effective manner of providing a central data warehouse, particularly via employing a tool that helps by analyzing existing data and reports from different business units. In accordance with at least one embodiment of the invention, such a tool analyzes the data model of an enterprise and proposes alternatives for building a new data warehouse. The tool, in accordance with at least one embodiment of the invention, models the problem of identifying fact/dimension attributes of a warehouse model as a graph cut on a Dependency Analysis Graph (DAG). The DAG is built using existing data models and the report generation scripts. The tool also uses the DAG for generation of ETL (Extract, Transform Load) scripts that can be used to populate the newly proposed data warehouse from data present in the existing schemas.

Abstract: A computer receives one or more files having configuration information that includes data that defines a plurality of stages of an extract, transform, and load (ETL) job, wherein the plurality of stages comprise a read stage that is preceded by a write stage, and wherein the read stage reads data from a source location, and wherein the data that is read or a modified version of the data that is read is being written by the write stage that writes data to the source location. The computer replaces the read stage with a decompressor stage. The computer replaces the write stage with a compressor stage. The computer executes the decompressor stage and compressor stage on a field-programmable gate array that is programmatically customized with data compression and data decompression functionality to enhance the performance of the ETL job.