Abstract: A scheme (300) for facilitating automated AR-based rendering with respect to expert guidance or assistance provided in a connected work environment based on contextualization is disclosed. In one aspect, a method comprises a worker (102) requiring assistance with respect to a given task generating (302) one or more suitable queries. Responsive thereto, real-time context data is gathered (304), which may comprise data relating to the work environment. A remote expert (118) may generate (306) appropriate guidance relevant to the task and assistance query, which may be rendered as AR content (308) for worker consumption by using an AR rendering module (112). The worker (102) consumes or uses the AR content for performing (310) an action with respect to the task. The performance of the worker (102) may be used as a measure to improve the AR rendering in automated fashion (312, 314) using one or more machine learning modules (114).

Abstract: An apparatus and method are described herein for training a similarity model that is used to predict similarity between item pairs. In one example, the apparatus is configured to train (build) the similarity model by using machine learning (e.g., a multivariate multiple liner regression process) that utilizes an independent variable including metadata (e.g., title, genre, writer, plot keywords) associated with training items, and two dependent variables including user contributed similarity scores for training item pairs, and collaborative filtering similarity scores for the training item pairs. Then, the apparatus uses the trained model to predict similarity between items.

Abstract: A Head-Mounted Display (HMD) is provided which comprises a camera configured to capture a video of a real-world scene with a first field-of-view (FoV), a network interface circuit configured to stream video to a receiving display device, and processing means which is operative to generate a 3D model of the real-world scene, and to generate a video from the 3D model using a second FoV which is wider than the first FoV. The processing means is further operative to estimate a motion of the camera, and to stream the generated video to the receiving display device if the estimated motion of the camera satisfies one or more conditions indicative of rapid motion, else stream the captured video to the receiving display device.

Abstract: The present disclosure relates to a method and an ad decision sever for determining personalized messages during a provisioning of digital content. The method comprises to provide a set of personalized messages candidates by executing a relevance algorithm, which matches message metadata and content metadata. The method continues with receiving calculated network costs for the set of personalized message candidates in order to determine a set of personalized messages, which are optimized with respect to the received network costs by applying a multivariate optimization algorithm.

Abstract: A method performed by an electronic device for requesting tiles relating to a viewport of an ongoing omnidirectional video stream is provided. The ongoing omnidirectional video stream is provided by a server to be displayed to a user of the electronic device. The electronic device predicts for an impending time period, a future head gaze of the user in relation to a current head gaze of the user, based on: A current head gaze relative to a position of shoulders of the user, a limitation of the head gaze of the user bounded by the shoulders position of the user, and a current eye gaze and eye movements of the user. The electronic device then sends a request to the server. The request requests tiles relating to the viewport for the impending time period, selected based on the predicted future head gaze of the user.

Abstract: A tag-based, user-directed media recommendation scheme is described herein. For example described herein is a recommender system (and method implemented by the recommender system) comprising: a recommendation engine configured to generate a recommendations list which includes a current set of recommendations for a user of a client device; and, a tag engine configured to: (1) receive the recommendations list; (2) obtain information about representative items associated with the user of the client device; (3) correlate the recommendations with tags; (4) correlate the representative items with the tags, (5) sort the tags into a tag list; and (6) provide the recommendations list and the tag list to the client device. The representative items include at least one of following: (1) an item previously purchased by the user; (2) an item previously watched by the user; (3) an item previously placed on a wish-list by the user; and (4) an item recommended by the recommendation engine for the user.

Abstract: A Head-Mounted Display (HMD) is provided which comprises a camera configured to capture a video of a real-world scene with a first field-of-view (FoV), a network interface circuit configured to stream video to a receiving display device, and processing means which is operative to generate a 3D model of the real-world scene, and to generate a video from the 3D model using a second FoV which is wider than the first FoV. The processing means is further operative to estimate a motion of the camera, and to stream the generated video to the receiving display device if the estimated motion of the camera satisfies one or more conditions indicative of rapid motion, else stream the captured video to the receiving display device.

Abstract: A scheme (300) for facilitating automated AR-based rendering with respect to expert guidance or assistance provided in a connected work environment based on contextualization is disclosed. In one aspect, a method comprises a worker (102) requiring assistance with respect to a given task generating (302) one or more suitable queries. Responsive thereto, real-time context data is gathered (304), which may comprise data relating to the work environment. A remote expert (118) may generate (306) appropriate guidance relevant to the task and assistance query, which may be rendered as AR content (308) for worker consumption by using an AR rendering module (112). The worker (102) consumes or uses the AR content for performing (310) an action with respect to the task. The performance of the worker (102) may be used as a measure to improve the AR rendering in automated fashion (312, 314) using one or more machine learning modules (114).

Abstract: The present disclosure relates to a method and an ad decision sever for determining personalized messages during a provisioning of digital content. The method comprises to provide a set of personalized messages candidates by executing a relevance algorithm, which matches message metadata and content metadata. The method continues with receiving calculated network costs for the set of personalized message candidates in order to determine a set of personalized messages, which are optimized with respect to the received network costs by applying a multivariate optimization algorithm.

Abstract: A tag-based, user-directed media recommendation scheme is described herein. For example described herein is a recommender system (and method implemented by the recommender system) comprising: a recommendation engine configured to generate a recommendations list which includes a current set of recommendations for a user of a client device; and, a tag engine configured to: (1) receive the recommendations list; (2) obtain information about representative items associated with the user of the client device; (3) correlate the recommendations with tags; (4) correlate the representative items with the tags, (5) sort the tags into a tag list; and (6) provide the recommendations list and the tag list to the client device. The representative items include at least one of following: (1) an item previously purchased by the user; (2) an item previously watched by the user; (3) an item previously placed on a wish-list by the user; and (4) an item recommended by the recommendation engine for the user.

Abstract: A system and method are described herein for improving content-based (CB) filtering to find other items similar to a particular item. For example, the system and method can operate to (1) collect ground truth from users, experts or non-experts for a subset of items in a library; (2) build similarity model(s) against the collected ground truth and evaluate this/these similarity model(s) for generalizability to all items in the library, and (3) build recommendations for users based on knowledge elicited from the users themselves, combined with the similarity scores in the built similarity model(s).

Abstract: A tag-based, user-directed media recommendation scheme is described herein. For example described herein is a recommender system (and method implemented by the recommender system) comprising: a recommendation engine configured to generate a recommendations list which includes a current set of recommendations for a user of a client device; and, a tag engine configured to: (1) receive the recommendations list; (2) obtain information about representative items associated with the user of the client device; (3) correlate the recommendations with tags; (4) correlate the representative items with the tags, (5) sort the tags into a tag list; and (6) provide the recommendations list and the tag list to the client device. The representative items include at least one of following: (1) an item previously purchased by the user; (2) an item previously watched by the user; (3) an item previously placed on a wish-list by the user; and (4) an item recommended by the recommendation engine for the user.

Abstract: An apparatus and method are described herein for training a similarity model that is used to predict similarity between item pairs. In one example, the apparatus is configured to train (build) the similarity model by using machine learning (e.g., a multivariate multiple liner regression process) that utilizes an independent variable including metadata (e.g., title, genre, writer, plot keywords) associated with training items, and two dependent variables including user contributed similarity scores for training item pairs, and collaborative filtering similarity scores for the training item pairs. Then, the apparatus uses the trained model to predict similarity between items.

Abstract: A tag-based, user-directed media recommendation scheme is described herein. For example described herein is a recommender system (and method implemented by the recommender system) comprising: a recommendation engine configured to generate a recommendations list which includes a current set of recommendations for a user of a client device; and, a tag engine configured to: (1) receive the recommendations list; (2) obtain information about representative items associated with the user of the client device; (3) correlate the recommendations with tags; (4) correlate the representative items with the tags, (5) sort the tags into a tag list; and (6) provide the recommendations list and the tag list to the client device. The representative items include at least one of following: (1) an item previously purchased by the user; (2) an item previously watched by the user; (3) an item previously placed on a wish-list by the user; and (4) an item recommended by the recommendation engine for the user.

Abstract: A system (e.g., content management system), a recommender system, a storage device, and various methods are described herein that improves with the aid of human (expert) judgment the online-learning for item-item similarity (e.g., movie-movie similarity).

Abstract: A system and method are described herein for improving content-based (CB) filtering to find other items similar to a particular item. For example, the system and method can operate to (1) collect ground truth from users, experts or non-experts for a subset of items in a library; (2) build similarity model(s) against the collected ground truth and evaluate this/these similarity model(s) for generalizability to all items in the library, and (3) build recommendations for users based on knowledge elicited from the users themselves, combined with the similarity scores in the built similarity model(s).