Knowledge sharing large language model collaborative optimization method and system

The invention relates to a knowledge sharing big language model collaborative optimization method and system, and the method achieves the purpose of data privacy protection through a knowledge sharing mode, and guarantees the semantic comprehension and generation capability of a local language model of each client through a multi-client collaborative optimization mode. Model training is distributed to a plurality of local clients through collaborative learning, centralized storage of centralized data is avoided, and the distributed learning mode enables the model to learn rich language knowledge from a massive corpus. According to the method, when the client carries out judgment, the distribution similarity of the server and the client is measured through the KL divergence, so that the difference between the server and the client is determined, and the purpose of personalized learning is achieved. 本申请涉及了一种知识共享的大语言模型协同优化方法和系统，方法通过知识共享的方式达到数据隐私保护的目的，通过多客户端协同优化的方式保证每一客户端的本地语言模型的语义理解与生成能力，通过协同学习将模型训练分布到多个本地的客户端上，