Abstract: 【Purpose/significance】Academic innovation opportunity discovery is an important way to determine the direction of innova⁃ tion and lead the development of innovation. Discovering academic innovation opportunities based on sentence generation models can promote the intelligent development of academic innovation opportunity discovery and promote scientific research innovation【. Method/ process】Taking management as the source discipline and information science as the target discipline, the literature in the core jour⁃ nals of management(CSSCI) is firstly clustered, and the subject words of each cluster are used as the starting point of the sentences, which are generated by using the intelligence corpus fine-tuning GPT-2 model. Design the innovation contribution index, calculate the innovation contribution score of the generated sentences, and regard the sentences with the highest innovation contribution score as academic innovation opportunities.Finally,the robustness of this paper's methodology is verified by ANOVA, and the validity of aca⁃ demic innovation opportunities from the perspectives of national policies and academic needs was verified.【Result/conclusion】 Through empirical research, the methodology of this paper accurately identifies academic innovation opportunities imported from man⁃ agement to information science,such as "accelerating the transfer of patented technology and enhancing the application value of pat⁃ ents" and "promoting the open circulation and utilisation of public data resources", etc. The analysis finds that the methodology of this paper can also be generalised to identify disruptive innovation opportunities in heterogeneous fields. It is found that the methodology of this paper can also be extended to the identification of disruptive innovation opportunities in heterogeneous fields【. Innovation/limita⁃ tion】To further enrich the method of discovering interdisciplinary innovation opportunities, a sentence generation model is introduced to discover academic innovation opportunities on the basis of new sentences. However, the corpus of this study is all in Chinese, and the innovation opportunities in foreign language literature are not studied, and the innovation opportunities discovered are not yet per⁃ fect.