Fecha de publicación: -- Fuente: WIPO "tomato" A system performs inference in conjunction with a machine-learned language model to predict a state of a receptacle that is located at a physical location and has sensors attached to the receptacle. The predicted state of the receptacle includes a predicted list of objects that are in a storage area of the receptacle, for example. The system may generate a first predicted state of the receptacle from receptacle sensor data of the sensors. A rule-based model may be used to generate the first predicted state, for example. The machine-learned language model may be used in conjunction with the rule-based model to also predict the state of the receptacle. The machine-learned language model better understands complex user behavior compared to the rule-based model resulting in increased accuracy of the predictions.