SDG Classification

SDG classification research focuses on automatically assigning scientific publications and company activities to the relevant UN Sustainable Development Goals (SDGs), facilitating impact assessment and resource allocation. Current efforts leverage machine learning, particularly deep learning methods and large language models (LLMs), often in combination with similarity measures and knowledge graphs, to improve the accuracy and efficiency of this classification. However, challenges remain regarding inherent biases in existing SDG classifications and the need for specialized models to overcome limitations of general-purpose LLMs, highlighting the importance of developing robust, transparent, and unbiased methodologies. This work has significant implications for monitoring SDG progress, informing policy decisions, and optimizing resource allocation across various sectors.