nai

Project Progress

AIRFrame analyzes published and unpublished documents to identify and visualize implicit relationships between astrobiology’s diverse constituent fields. The main goal of the AIRFrame project is to allow researchers and the public to discover related information from different disciplines.

During the reporting period, AIRFrame researchers harvested and analyzed over four thousand astrobiology publications and cited works over ten years, and compared the results of several bibliometric and computational document clustering methods to provide a measurement of the interdisciplinarity of astrobiology researchers, teams and papers, to reveal implicit connections between them, and to problematize the assignment of papers to rigid topical categories in scientific databases. By analyzing not just the text of astrobiology papers but also the abstracts of the papers they cite, we hope to transcend some of the semantic barriers between researchers from different fields, and develop a robust method to reveal implicit connections across all NAI teams, and interdisciplinary scientific collaborations more generally. We are also developing a new algorithm for selecting which words in a corpus of astrobiology documents contribute the most to the mutual information between the words and the documents, to generate data-driven topic labels for astrobiology document subsets, and evaluate the extent to which NAI goals and subgoals are being addressed by NAI research output.