Abstract:

In this paper we take military terminology as the research object and, in conjunction with military corpora, employ a pre-trained language model that has achieved groundbreaking progress in natural language processing in recent years to realize the automatic extraction of military terms. The foundational work of the research includes the construction of a large-scale military corpus and the design of the extraction model. Using ChatGLM as the base model, the extraction of new military terms is realized through the “pre-training” and “fine-tuning” phases. The pre-training phase involves pre-training the model on a large-scale unlabeled military corpus to acquire military terminology information from the corpus.

Key words: military terminology, pre-trained language model, term extraction