Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating video anchors for a video. In one aspect, a method includes obtaining key moment identifiers for a video, where each key moment identifier includes a time index value specifying a playback time in the video, and is indicative subject matter of the video that has been determined to meet one or more interest criteria that define salient topics within the video. For each key moment identifier, a video anchor is generated, where each video anchor indicates a playback time for the video, and may include an image of a frame that occurs near the playback time. Upon a selection of the video anchor by the user, an instruction in the video anchor causes a video player to begin playback of the video at the playback time specified by the video anchor.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, are provided, for generating video anchors for a video. In one aspect, a method includes obtaining key moment identifiers for a video, where each key moment identifier includes a time index value specifying a playback time in the video, and is indicative subject matter of the video that has been determined to meet one or more interest criteria that define salient topics within the video. For each key moment identifier, a video anchor is generated, where each video anchor indicates a playback time for the video, and may include an image of a frame that occurs near the playback time. Upon a selection of the video anchor by the user, an instruction in the video anchor causes a video player to begin playback of the video at the playback time specified by the video anchor.

Abstract: In one aspect, a method includes obtaining videos and for each video: obtaining a set of anchors for the video, each anchor beginning at the playback time and including anchor text; identifying, from text generated from audio of the video, a set of entities specified in the text, wherein each entity in the set of entities is associated with a times stamp at which the entity is mentioned; determining, by a language model and from the text generated from the audio of the video, an importance value for each entity; for a subset of the videos, receiving rater data that describes, for each anchor, the accuracy of the anchor text in describing subject matter of the video; and training, using the human rater data, the importance values, the text, and the set of entities, an anchor model that predicts an entity label for an anchor for a video.

Abstract: In one aspect, a method includes obtaining videos and for each video: obtaining a set of anchors for the video, each anchor beginning at the playback time and including anchor text; identifying, from text generated from audio of the video, a set of entities specified in the text, wherein each entity in the set of entities is associated with a times stamp at which the entity is mentioned; determining, by a language model and from the text generated from the audio of the video, an importance value for each entity; for a subset of the videos, receiving rater data that describes, for each anchor, the accuracy of the anchor text in describing subject matter of the video; and training, using the human rater data, the importance values, the text, and the set of entities, an anchor model that predicts an entity label for an anchor for a video.

Abstract: In one aspect, a method includes obtaining videos and for each video: obtaining a set of anchors for the video, each anchor beginning at the playback time and including anchor text; identifying, from text generated from audio of the video, a set of entities specified in the text, wherein each entity in the set of entities is associated with a times stamp at which the entity is mentioned; determining, by a language model and from the text generated from the audio of the video, an importance value for each entity; for a subset of the videos, receiving rater data that describes, for each anchor, the accuracy of the anchor text in describing subject matter of the video; and training, using the human rater data, the importance values, the text, and the set of entities, an anchor model that predicts an entity label for an anchor for a video.