As an emerging technology in the field of artificial intelligence (AI), graph neural networks (GNNs) are deep learning models designed to process graph-structured data. Currently, GNNs are effective at capturing relationships between nodes and edges in data, but often overlook higher-order, complex connections. To address this challenge, a research team at The Hong Kong Polytechnic University (PolyU) has developed a new heterogeneous graph attention network, revolutionising the modelling of complex relationships in graph-structured data. This innovation is poised to break through AI application limitations in fields such as neuroscience, logistics, computer vision and biology.

To achieve a deep understanding of the CMAS corrosion mechanism and lifetime prediction of high-performance thermal/environmental barrier coating materials, the (Er_1/4Y_1/4Lu_1/4Yb_1/4)₂Si₂O₇ and (Er_1/6Tm_1/6Y_1/15Gd_1/15Lu_4/15Yb_4/15)₂Si₂O₇ high-entropy rare-earth disilicates designed in this study exhibit approximately 70% reduction in CMAS corrosion depth compared to their single-principal-component counterparts, demonstrating excellent CMAS corrosion resistance. The research further reveals that lattice distortion induced by multi-cation doping can inhibit the penetration of CMAS melt, while large-radius rare-earth ions reduce the corrosion activity by consuming Ca²⁺ in the melt. Additionally, it elucidates the temperature-dependent transition of corrosion mechanisms—dominantly governed by thermodynamics–kinetics competition at 1300 °C, whereas shifting to a dissolution–reprecipitation mechanism at 1500 °C. On this basis, an extended Kalman filter model incorporating physical mechanisms was developed for the first time, enabling high-precision prediction of long-term corrosion depth and rate, thereby providing a reliable tool for coating lifetime assessment.