Abstract: Provided is a method of calculating an optimal temperature of a furnace, the method including: acquiring a plurality of appropriate candidate temperatures of the furnace from first input data using a trained optimal temperature model; acquiring predictive property values of the material from second input data using a trained property prediction model; acquiring one or more sampled appropriate candidate temperatures from the plurality of appropriate candidate temperatures based on a result of comparing a target property value with the predictive property values; acquiring heat information of a case of operating the furnace at the sampled appropriate candidate temperatures, acquiring one or more appropriate temperatures from the sampled appropriate candidate temperatures based on a result of comparing the heat information; and transmitting a request to set the acquired optimal temperature to a set temperature value of the furnace.

Abstract: Disclosed is an electronic device for implementing a temperature prediction and control system.

Abstract: Disclosed is an electronic device for implementing an industrial process prediction and control system. The electronic device includes one or more processors configured to perform predicting on a calorific value of recycled fuel and a temperature of a preheating chamber in a cement manufacturing apparatus using the trained first neural network model and the trained second neural network model based on the process information including fuel input information of a cement manufacturing apparatus, and controlling input fuel for cement based on this prediction.

Abstract: Disclosed is a method and apparatus for extracting data in a deep learning model. The method includes receiving an input query, determining a first decision boundary set being a subset of a decision boundary set corresponding to a target layer of the deep learning model, extracting a decision region including the input query based on the first decision boundary set, and extracting data included in the decision region.

Abstract: Disclosed is a method and apparatus for extracting data in a deep learning model. The method includes receiving an input query, determining a first decision boundary set being a subset of a decision boundary set corresponding to a target layer of the deep learning model, extracting a decision region including the input query based on the first decision boundary set, and extracting data included in the decision region.