Abstract: Embodiments are disclosed for crash detection on one or more mobile devices (e.g., smartwatch and/or smartphone). In some embodiments, a method comprises: detecting, with at least one processor, a crash event on a crash device; extracting, with the at least one processor, multimodal features from sensor data generated by multiple sensing modalities of the crash device; computing, with the at least one processor, a plurality of crash decisions based on a plurality of machine learning models applied to the multimodal features; and determining, with the at least one processor, that a severe vehicle crash has occurred involving the crash device based on the plurality of crash decisions and a severity model.

Abstract: One variation of a method includes: serving a class—spanning a duration of time, including an audio track of a trainer narrating an exercise routine of a first type, and associated with a first difficulty level—to a user; if the user exits the class within a first time window during replay of the class, presenting a first list of classes—spanning durations approximating the duration and including audio tracks of trainers narrating exercise routines of the first type—to the user; and, if the user exits the class within a second time window succeeding the first time window, serving a prompt to select a second class—from a second list of classes—spanning durations less than the first duration and including audio tracks of trainers narrating exercise routines of difficulty levels less than the first difficulty level—for completion during a remainder of the exercise session to the user.