Understanding the origin and evolution of molecular complexity in the interstellar medium is one central problem in astrophysics and astrochemistry. This field has advanced through traditional approaches, including chemical reaction networks, rate-equations, gas-grain modeling, quantum mechanical calculations, and radiative-transfer-based spectral analysis. These methods have provided the essential foundation for interpreting how molecules form and evolve in different astrophysical environments. However, with the rapid growth of modern observational data from facilities such as ALMA and JWST, the field is increasingly challenged by severe line confusion, incomplete reaction networks, and the difficulty of extracting robust physical and chemical information from complex spectra.
In this talk, I will present our recent efforts to build on traditional modeling and incorporate AI-enabled methods for exploring molecular complexity in the interstellar medium. This framework combines physics-based modeling with new data-driven and machine-learning tools for reaction-network expansion, multiphase chemical evolution, and automated spectral-line identification. I will discuss how recent developments such as ChemiVerse, GraSSCoL, and Spectuner can be integrated with traditional modeling strategies to build a more scalable and interconnected framework linking reaction pathways, abundance evolution, synthetic spectra, and observational inference. By combining traditional modeling with AI-enabled methods, we aim to improve the efficiency and reliability of molecular identification and to support more systematic studies of molecular complexity in the interstellar medium. The broader goal is to develop a more predictive framework for understanding how complex, and potentially prebiotic, molecules form and evolve in space.
