A method, a system, and a non-transitory storage media for determining agent importance prediction for autonomous driving. Input features associated with agents present in an environment surrounding a vehicle are processed. Output features associated with the agents are determined based on the input features. An importance of each agent is predicted using the output features. One or more movements of the vehicle are determined based on the predicted agent importance.

A system and method for model predictive control of a process for removing dissolved oxygen (DO) from seawater to produce treated seawater having less than a prescribed DO concentration and a prescribed pH is disclosed. The model predictive control system includes a machine learning (ML) module for calculating, based on the values of operational input parameters, a predicted DO concentration and a predicted pH of the treated seawater for a future point in time. An ML-based control module is configured to determine, based on the predicted DO concentration, predicted pH and the input parameters, settings for adjusting controllable operational input parameters that serve to change the DO concentration or pH of the treated seawater. The control system monitors DO and pH during operation to dynamically update the DO and pH predictions, and adaptively update system settings to produce treated seawater having less than the prescribed DO concentration and pH.

A control device executes a step of starting a computation processing of a prediction model; a step of computing a remaining processing time until the computation processing is completed after starting the computation processing of the prediction model; a step of determining whether the determination of the command value based on an output obtained from the prediction model is made within a control timing for controlling the operation of manufacturing by the manufacturing device, on the basis of a computed remaining processing time; and a step of stopping, when it is determined that the determination of the command value is not made within the control timing, the computation processing of the prediction model, determining the command value on the basis of a value of an intermediate result of the computation processing, and controlling the operation of the manufacturing device on the basis of the determined command value.

The present invention reduces the probability of malfunction occurrence when performing predictive control of a device being controlled. In this control device of one aspect of the present invention, a prediction model for a control variable is used to calculate a prediction value from a measured value of the control variable, and a desired command value of the control variable is determined by correcting a desired basic value in accordance with the calculated prediction value. The degree of correction is determined on the basis of weight. The control device controls the operation of the device being controlled according to the determined desired command value. The control device assesses whether the device being controlled is operated appropriately on the basis of monitoring data relating to the operation result of the device being controlled, and optimizes the weight of the correction to make appropriate control possible based on the assessment result.

A dataset regarding a plurality of applications is obtained. A set of parameters is determined from the dataset, comprising at least a sample performance trajectory, a risk factor, and a performance outcome. A maximum likelihood of each performance outcome is determined using a likelihood function, the likelihood function being a mixture model of a trajectory model and an outcome model. The set of parameters is updated according to the maximum likelihood of each performance outcome. A performance trajectory model is built according to the updated set of parameters. The plurality of applications is then grouped into subgroups according to the performance trajectory model, each subgroup containing one or more applications, and each of the one or more applications in a given subgroup having a same or similar trajectory to each other. An alert associated with the applications in at least one of subgroups may be generated.

A method for predicting a path of a specific occupant of a number of occupants including receiving user access data, the user access data including a user identifier, an access time, and an access location, generating, a first model describing general sequences of access events associated with the number of occupants and a frequency of each of the general sequences, generating a second model describing specific sequences of access events associated with the specific occupant and a frequency of each of the specific sequences, and generating a path prediction model based on the first and second models, the path prediction model including a weighted score for each of the number of access control points, the weighted score associated with a probability the specific occupant accesses the access control point based on a last accessed access control point.

A method of remotely activating a vehicle system of a vehicle, the vehicle being part of a vehicle-share enterprise, the method including generating a control signal based on session parameters associated with a vehicle-sharing session, preconditioning parameters associated with the vehicle-sharing session, and sensor data associated with the vehicle-sharing session, where the sensor data includes vehicle sensor data from one or more sensors of the vehicle associated with the vehicle-sharing session, client device sensor data from one or more sensors of a client device, or a combination thereof. The method includes, prior to initiating the vehicle-sharing session, broadcasting the control signal to the vehicle, where the control signal is configured to selectively control the vehicle system.

A method may include obtaining first ambient condition data corresponding to a first building in a first location. The method may further include obtaining a set of ice dam models. The method may further include predicting, based at least in part on the set of ice dam models, an ice dam formation on the first building. The method may further include obtaining a heating profile. The heating profile may be based at least in part on the first ambient condition data. The method may further include adjusting, based on the heating profile, a heating device of the first building.

A building management system including building equipment operable to affect a variable state or condition of a building. The building management system includes a controller including a processing circuit. The processing circuit is configured to obtain an energy prediction model (EPM) for predicting energy requirements over time. The processing circuit is configured to monitor one or more triggering events to determine if the EPM should be retrained. The processing circuit is configured to, in response to detecting that a triggering event has occurred, identify updated values of one or more hyper-parameters of the EPM. The processing circuit is configured to operate the building equipment based on the EPM.

A method organizes fibers. A plurality of fibers is received into a first assembly. An initial sequence of the plurality of fibers in the first assembly is obtained. A set of key combinations is identified from the initial sequence and a predetermined sequence. A second assembly is slid across the first assembly. The set of key combinations is actuated to move the plurality of fibers from the first assembly to the second assembly and order the plurality of fibers in the second assembly in the predetermined sequence.