Research on Training Scenario Assisted Design Method Based on Domain-Adapted Language Model

In recent years, the generative artificial intelligence technology has developed rapidly, showing great application potential in the field of training. Training scenario design requires familiarity with relevant training concepts, equipment performance, force deployment, force action and other information, and it is difficult and inefficient to rely solely on manual work. In view of this issue, we propose training scenario assisted design method based on domain-adapted language model. First, we build the training scenario question and answer database to fine-tune the large language model, then use chain of thought and retrieval-augmented to accelerate the process of training scenario design. The experimental results show that the proposed method can quickly generate the training scenario summary document and assist the relevant personnel to complete the simulation scenario design efficiently.