Abstract: A machine learning (ML) computer model is trained on a dual-modal input sequence comprising target content data and an input user query specifying attributes for locating a portion within the target content data. The ML computer model is trained to process the dual-modal input sequence and output an output text specifying start and end indices, within the target content data, corresponding to the portion. Runtime content data, and a user spoken query, are received and the runtime content data and user spoken query are processed via the trained ML computer model to generate an output text specifying a predicted beginning index and predicted ending index within the runtime content data. The portion of the runtime content data corresponding to a range of runtime content data bounded by the beginning index and ending index, inclusively, is retrieved and stored for later retrieval and processing.

Abstract: A computer adjusts a set of presentation materials, comprising. The computer receives, from a source available to the computer, an initial set of presentation materials. The computer, in response to receiving the initial set of presentation materials, determines a reference presentation duration associated with initial set of presentation materials. The computer receives a target presentation duration from a target duration source available to the computer. The computer determines a presentation conversion value representing, at least in part, a ratio of the target presentation duration to the reference presentation duration. The computer applies a Machine Learning (ML) refactoring routine to revise the initial set of presentation materials in accordance, at least partially, with the presentation conversion value, thereby generating a refactored set of presentation materials having a revised conveyance duration substantially the same as the target duration.