Samsung patents translation check that swaps bad text for known good examples

1. An electronic device comprising: a communicator; at least one memory storing at least one instruction and storing learning data comprising a plurality of encoding vectors; and at least one processor configured to access the at least one memory and to execute the at least one instruction, wherein the at least one instruction, when executed by the at least one processor, causes the electronic device to: based on acquiring information on an original text of a first language, input the information on the original text into a first encoder and a second encoder, acquire as an output of the first encoder a third encoding vector comprising information for acquiring a translation text corresponding to the original text, and acquire as an output of the second encoder a first encoding vector comprising semantic information included in the original text and semantic information included in the third encoding vector, input the first encoding vector into a discriminator and acquire as an output of the discriminator a probability value indicating a probability that the first encoding vector corresponds to a normal translation, based on the probability value being greater than or equal to a predetermined first threshold value, control the communicator to transmit the first encoding vector to an external device comprising a decoder configured to acquire a translation text of a second language corresponding to the original text, based on the probability value being smaller than the predetermined first threshold value, acquire a similarity value for each of the plurality of encoding vectors, wherein each similarity value indicates a degree of similarity between a respective encoding vector of the plurality of encoding vectors and the first encoding vector, identify a second encoding vector among the plurality of encoding vectors having the highest similarity value, and control the communicator to transmit the identified second encoding vector to the external device, store the third encoding vector and the first encoding vector in the at least one memory as part of the learning data, train the second encoder based on the learning data which includes the third encoding vector and the first encoding vector, and repeat the foregoing operations on a second original text using the first encoder and the trained second encoder.