Systems and methods are described herein for curating content that follows a narrative structure. A narrative structure comprises narrative portions that have a defined order. Signature analysis of known content that follows the narrative structure is used to train machine learning models for the narrative structure and the narrative portions that make up the narrative structure. Signature analysis of candidate content segments, along with machine learning models for the narrative portions, are used to identify candidate content segments that match the respective narrative portions. A candidate playlist is generated of the identified candidate content segments in the defined order. In one embodiment, the machine learning model for the narrative structure is used to validate the generated playlist.

Methods, systems, and media for indicating viewership of a video are provided. In some embodiments, the method comprises: identifying a video; identifying a first group of users; determining an affinity score for each user with the identified video; receiving a request to present a page that includes a representation of the video; identifying a second group of users connected to the first user associated with the user device; determining a viewership status of each user in the second group of users corresponding to the video; identifying a subset of users in the second group of users based at least in part on the viewership status; and causing groups of indicators to be presented on the user device, wherein each indicator in the groups of indicators represents the viewership status of the user, and wherein the indicators are presented on the requested page in connection with the representation of the video.

Systems and methods for, while a user is consuming a first media asset, generating for the user a recommendation of a second media asset based on a viewing history of the user. In some aspects, the systems and methods determine that a user is consuming only video of a first media asset, determine a first category for the first media asset, and retrieve a viewing history corresponding to the user. The viewing history comprises media assets consumed simultaneously with a media asset corresponding to the first category. The systems and methods select a second category corresponding to a highest number of media assets in the viewing history, select a second media asset based on the second category. Alternatively, the user can select the second media asset manually. The systems and methods generate for output the audio of the second media asset for simultaneous consumption with the video of the first media asset.

A multimedia information processing method includes: parsing multimedia information to separate an audio from the multimedia information; converting the audio to obtain a mel spectrogram corresponding to the audio; determining, according to the mel spectrogram corresponding to the audio, an audio feature vector corresponding to the audio; and determining, based on an audio feature vector corresponding to a source audio in source multimedia information and an audio feature vector corresponding to a target audio in target multimedia information, a similarity between the target multimedia information and the source multimedia information.

Systems and techniques are provided for automation and recommendation based on device control protocols. HDMI-CEC device control data may be received from a connected electronic device that may be connected to an electronic display device. The HDMI-CEC device control data may be based on a HDMI-CEC device control signal from the electronic display device. The system may generate a control signal for a device. The control signal may be sent to the device for implementation.

An information processing apparatus that processes information on the basis of a gaze degree of a user who views content is provided.

An information processing apparatus includes: an estimation unit that estimates a gaze degree of a user who views content; an acquisition unit that acquires related information to content recommended to the user; and a control unit that controls a user interface that presents the related information on the basis of an estimation result of the gaze degree. The acquisition unit acquires the related information by using an artificial intelligence model that has learned a causal relationship between information on a user and content in which a user shows interest.

There is described an apparatus for providing a recommendation to a first user, wherein the apparatus comprises: a processor arranged to: determine, at a first time, a first set of vectors for a plurality of users, including a vector associated with the first user; determine a set of clusters based on the first set of vectors; detect, at a second time, an event associated with the first user; determine an updated vector for the first user in dependence on the event; place the first user into a cluster of the set of clusters based on the updated vector; and determine a recommended item of content for the first user in dependence on the cluster; and a user interface and/or communication interface arranged to: output the recommendation and/or the recommended item to the first user.

A media stream is delivered to a searcher by identifying a desirable emotional state for the searcher; identifying a current emotional state of the searcher; and estimating a target emotional trajectory that begins with the current emotional state and concludes with the desirable emotional state. Then the target emotional trajectory is matched to an aggregate emotional trajectory of the media stream; the media stream is recommended to the searcher in response to the matching; and a bit stream of the media stream is rendered to the searcher in response to the searcher's acceptance of the recommendation.

An example method involves accessing a playlist defining a sequence of media content items including a first media content item and a second media content item; retrieving, from one or more server devices, first data representing the media content items of the received playlist, and using the retrieved first data to playout the media content items in accordance with the sequence defined by the playlist; determining a status of an event, and using the determined status of the event as a basis to determine a condition; making a determination that the determined condition has been satisfied; and responsive to making the determination that the determined condition has been satisfied, (i) retrieving second data representing a media content item related to the event, and (ii) using the retrieved second data to playout the media content item related to the event in between playing out the first and second media content items.

Systems and methods relate to display of insights, including a method including: causing a user display to display a graphical user interface (GUI) that enables a user to select a time range and a single digital display as a locale; and causing, based on the user selecting the time range and the single digital display as the locale, the user display to display, on the GUI, a list of a plurality of media content that have been displayed on the single digital display, each of the plurality of media content being ordered according to analysis insight for the single digital display selected and the time range selected. The analysis insight may include at least impression counts that are counts of views by people of the plurality of media content on one or more digital displays associated with the locale.