Research Progresses\nand Applications of Knowledge\nGraph Embedding Technique in Chemistry

A knowledge graph (KG) is a technique for modeling entities\nand\ntheir interrelations. Knowledge graph embedding (KGE) translates these\nentities and relationships into a continuous vector space to facilitate\ndense and efficient representations. In the domain of chemistry, applying\nKG and KGE techniques integrates heterogeneous chemical information\ninto a coherent and user-friendly framework, enhances the representation\nof chemical data features, and is beneficial for downstream tasks,\nsuch as chemical property prediction. This paper begins with a comprehensive\nreview of classical and contemporary KGE methodologies, including\ndistance-based models, semantic matching models, and neural network-based\napproaches. We then catalogue the primary databases employed in chemistry\nand biochemistry that furnish the KGs with essential chemical data.\nSubsequently, we explore the latest applications of KG and KGE in\nchemistry, focusing on risk assessment, property prediction, and drug\ndiscovery. Finally, we discuss the current challenges to KG and KGE\ntechniques and provide a perspective on their potential future developments.