Machine learning maps Hindu and Buddhist deities
Summary and headline written by AI from the source article. How we work
Researchers developed a system to compare 196 Hindu and Vajrayana Buddhist deities using machine learning, seeking to quantify similarities and differences in their visual forms and theological roles.
The system uses a “dual-matrix computational architecture” to analyze both physical appearance and the meanings associated with each deity. A “Cardinality Weighting” algorithm helped measure visual differences, while a Large Language Model created semantic expansions to represent theological functions, avoiding circular reasoning.
The analysis revealed instances of “iconographic camouflage,” where different appearances hide shared functions across the two traditions. The researchers also modeled the “Atin Effect,” a cognitive bias, and used it to test the machine learning process. Their work showed that certain high-level deities, like the Hindu Chinnamasta and the Buddhist Chinnamunda, occupy nearly the same mathematical space when analyzed by both visual and semantic criteria.
The researcher made the system openly available, intending it as a tool for scholars in Digital Humanities and comparative theology to identify structural connections within cultural information.
