Emotions come from the long-term evolution of daily language, DeepSeek-like generative artificial intelligence can gain insights into the different language habits and multicultural backgrounds implied in users’ utterances, thus providing a new source of motivation and strong technical support for the continuation of the vitality of multicultural civilization. In this paper, we take advantage of the core technology of DeepSeek tool to calculate user’s emotions with its large language model, and further propose a deep learning-based emotion dialog generation model. The emotional dialog generation model is combined with the bionic intelligent technology development of “bionic intelligence” to design the emotional expression of humanoid robots. Analyze the performance of the proposed emotional dialog generation model based on recurrent convolutional network on different indexes, and analyze the feasibility of DeepSeek accessing multi-channel apps for effective emotion recognition. The humanoid robot is subjected to experiments on multi-language emotion expression to analyze the possibility of its multi-language communication. DeepSeek is able to distinguish the emotion expression and emotion intention of different app users by accessing to multiple apps. The humanoid robot’s multiple emotional expressions can be universally recognized and can be used as an innovative way of development in multicultural communication.
