Executive Summary : | Multi-chromophore interactions are crucial for excited state energy and electron transfer processes in biological functions like light harvesting and photosynthesis. They are also important in energy research, particularly in the context of singlet fission phenomena, which have been used to create hybrid solar cell materials. Accurate electronic structure calculations of excited states are required for in-depth understanding and formulating design principles. However, these calculations can be computationally challenging and sometimes impossible for large systems. To address this, a combination of electronic structure methods and machine learning is proposed. This will lead to faster computational methods and machine learning-based workflows, enabling high throughput screening of important materials. The research aims to develop machine learning-based methods that depend on physics-based wavefunction ansatz, retaining the interpretability of wavefunction-based methods. The project will also develop a high-throughput workflow, allowing for the definition of suitable design principles for singlet fission materials. |