This review discussed the dilemma of small data faced by materials machine learning. First, we analyzed the limitations brought by small data. Then, the workflow of materials machine learning has been introduced. Next, the methods of dealing with small data were introduced, including data extraction from publications, materials database construction, high-throughput computations and experiments from the data source level; modeling algorithms for small data and imbalanced learning from the algorithm level; active learning and transfer learning from the machine learning strategy level. Finally, the future directions for small data machine learning in materials science were proposed.

The development of materials is one of the driving forces to accelerate modern scientific progress and technological innovation. Machine learning (ML) technology is rapidly developed in many fields and opening blueprints for the discovery and rational design of materials. In this review, we retrospected the latest applications of ML in assisting perovskites discovery. First, the development tendency of ML in perovskite materials publications in recent years was organized and analyzed. Second, the workflow of ML in perovskites discovery was introduced. Then the applications of ML in various properties of inorganic perovskites, hybrid organic–inorganic perovskites and double perovskites were briefly reviewed. In the end, we put forward suggestions on the future development prospects of ML in the field of perovskite materials.

As machine learning (ML) continues to advance in the field of materials science, the variation in strategies for the same steps of the ML workflow becomes increasingly significant. These details can have a substantial impact on results, yet they have not received the deserved attention. This review aims to explore the various strategies available for each detail within the general workflow of materials ML. Firstly, the general workflow of materials ML will be introduced to help readers gain an understanding of potential details. Subsequently, different strategies of details within each step of the workflow will be presented through state-of-the-art case studies. The potential outcomes associated with choosing different strategies to details will be explored. Following this, suitable strategies for details will be recommended based on distinct application scenarios. Finally, directions for the future development of materials ML concerning details will be proposed. Through these discussions, we aspire to offer a comprehensive understanding of the nuances in details within materials ML. This will serve as valuable reference and guidance for researchers in both materials science and ML.

We presented an effective and universal strategy for the improvement of luminophore’s solid-state emission, i.e., macrocyclization-induced emission enhancement (MIEE), by linking luminophores through C(sp³) bridges to give a macrocycle. Benzothiadiazole-based macrocycle (BT-LC) has been synthesized by a one-step condensation of the monomer 4,7-bis(2,4-dimethoxyphenyl)−2,1,3-benzothiadiazole (BT-M) with paraformaldehyde, catalyzed by Lewis acid. In comparison with the monomer, macrocycle BT-LC produces much more intense fluorescence in the solid state (Φ_PL = 99%) and exhibits better device performance in the application of OLEDs. Single-crystal analysis and theoretical simulations reveal that the monomer can return to the ground state through a minimum energy crossing point (MECP_S1/S0), resulting in the decrease of fluorescence efficiency. For the macrocycle, its inherent structural rigidity prohibits this non-radiative relaxation process and promotes the radiative relaxation, therefore emitting intense fluorescence. More significantly, MIEE strategy has good universality that several macrocycles with different luminophores also display emission improvement.