A study led by Jun Cheng from Xiamen University and collaborators introduces a new workflow for recommending relay catalysis pathways. The workflow uses large language models (LLMs) to extract and organize catalytic reaction data, and combines this with a self-built catalysis knowledge graph (Cat-KG). The system automatically filters and recommends high-quality, traceable multi-step pathways, helping researchers design catalytic reactions more efficiently.

Inverse lithography technology (ILT) is driving transformative innovations in semiconductor patterning processes. This paper reviews the evolution of ILT, providing an analysis of the applications in semiconductor manufacturing. In recent years, artificial intelligence (AI) has introduced breakthroughs for ILT, such as convolutional neural networks, generative adversarial networks, and model-driven deep learning, demonstrating potential in large-scale integrated circuit design and fabrication. This paper discusses future directions for ILT, which is expected to provide insights into semiconductor industry development.