The project proposes a new methodology for screening, prediction and validation of photoactive perovskite materials using machine learning techniques (ML). In the past few years, solar cells based on hybrid perovskite materials have shown impressive values of photoconversion efficiencies (PCEs), to date reaching 25.2%, with options for further enhancement. However, due to the huge number of possible structural and compositional configurations, optimizing the perovskite materials for stability and solar cell PCE by exhaustive numerical calculations or large scale synthesis is not feasible. Instead, the ML techniques can provide the necessary framework for a guided search. Using high throughput density functional theory (DFT) calculations, a database containing opto-electronic properties of interest shall be first assembled. Then, the ML scheme shall be implemented using artificial neural networks (ANNs), which already provided successful predictions in other condensed matter systems. They will primarily use the theoretical data as well as feedback from experiments. The selected candidates shall be synthesized and perovskite solar cells shall be fabricated as final products. The aim is to optimize the absorption spectra of the perovskite materials in order to increase the solar cell PCE and to enhance their stability. The coordinator team (NIPNE) will be focused on the development of the DFT-ML scheme, based on prior experience with first-principles calculations and ANN based methods for the prediction of the electronic gaps. The partner team (NIMP) will perform the synthesis of perovskite materials and fabrication of PSCs, based on extensive expertise accumulated during the PERPHECT project, where record PCEs were achieved. Relating two key elements (ML techniques and perovskite materials) the project is expected to have a large impact in material engineering and can reshape the current approaches for investigating new materials.