Abstract: Systems and methods for training machine learning models over labeled and unlabeled datasets are provided. Labels are assigned to unlabeled data by selecting a labeling approach, such as active learning or semi-supervised learning, based on uncertainty in the model's predictions. The selection of the labeling approach may be varied over the course of training, e.g. so that unlabeled dataset samples with progressively more uncertain predictions are pseudo-labeled via semi-supervised learning rather than with active learning, thereby reducing the load on the oracle and recognizing the increasing confidence in the model's overall calibration as training progresses.

Abstract: Methods, systems, and computer-readable media for identifying intentions related to a user's present intention are provided. The user's present intention may be determined by mapping the user's current actions to a present intent. A related intent may be determined using an intent chain. The intent chain may be generated by analyzing patterns of user activities and mapping the patterns of activities to the underlying intents.

Abstract: Methods, systems, and computer-readable media for identifying intentions related to a user's present intention are provided. The user's present intention may be determined by mapping the user's current actions to a present intent. A related intent may be determined using an intent chain. The intent chain may be generated by analyzing patterns of user activities and mapping the patterns of activities to the underlying intents.