Using Machine Learning Language Models to Generate Innovation Knowledge Graphs for Patent Mining

To explore and understand the state-of-the-art innovations in any given domain, researchers often need to study many domain patents synthesize their knowledge content. This provides a smart patent graph generation system, adopting machine learning (ML) natural language modeling approach, help grasp by generating deep graphs. research focuses on converting chemical utility patents, consisting of chemistries processes, into summarized The methods are two parts, i.e., visualization processes patents’ most relevant paragraphs domain-specific collection texts. ML algorithms, including ALBERT for text vectorization, Sentence-BERT sentence classification, KeyBERT keyword extraction, adopted. These models trained tested case using 879 carbon capture domain. results demonstrate that average retention rate summary graphs five clustered texts exceeds 80%. proposed approach is novel proven be reliable graphical representation.