Optimizing Hyperparameters of Multi-round Dialogue Model Based on Multi-objective Optimization Algorithm

Dialogue models have extensive applications and attracted significant attention. However, in the field of hyperparameter optimization, previous methods often face challenges such as prolonged processing time and low accuracy. This study explores a method for optimizing hyperparameters of multi-round dialogue models based on a multi-objective optimization algorithm. Inspired by the evolutionary laws in nature. It proposes a multi-objective evolutionary algorithm capable of dynamically allocating computational resources. It can optimize the hyperparameters of multi-round dialogue models, thereby enhancing the model’s accuracy. A highly accurate multi-turn dialogue system can quickly complete the tedious work for people, thereby improving people’s quality of life. Compared with the existing work, our method demonstrates shorter processing time and higher accuracy via experiments.