Abstract:[Purpose/significance] From the perspective of causality,this paper proposes to excavate the logical chain of scientific and technological innovation from scientific events and discover potential innovation opportunities[Method/process] Firstly,the causal events with triplet structure were extracted from the scientific events,and then the events were globally correlated to build a causal networkThen,based on causal network,innovation path discovery based on semantic causality and innovation opportunity discovery based on representational causality are proposed respectivelyThrough network analysis,path calculation,feature representation,link prediction,and other methods to excavate the significant and strong continuity of scientific events and their evolution path in the process of scientific and technological innovation;and identify potential innovation elements with causal logic[Result/conclusion] The analysis shows that artificial intelligence and intelligent manufacturing play a very important role in scientific and technological innovation;astronomy,energy materials and other fields have stronger continuity in scientific innovation and play a stable and basic role;the humanities is not active in scientific research,so it needs to be closely combined with other disciplines in the futureIn addition,through the prediction of potential innovation opportunities,it is found that there is a potential causal relationship between global climate change and many aspects of scientific and technological development,and we should continue to pay attention to climate issues in the futureThis paper transforms the path discovery and element identification of scientific events into the causal network,excavates the logical chain of scientific and technological innovation through modeling,calculation and prediction,and provides practical method support for the scheme through empirical research,which can provide reference for the development of scientific innovation and the strategic layout of science and technology,which is effective and feasible