Audio Tagging

Audio tagging involves automatically assigning descriptive labels to audio segments, aiming to improve content organization, accessibility, and analysis. Current research focuses on developing efficient and accurate models, exploring architectures like transformers, graph neural networks, and CNNs, often incorporating techniques like knowledge distillation to reduce computational demands while maintaining performance. These advancements are impacting various fields, from enhancing music streaming services and broadcasting workflows to improving audio indexing and retrieval in large-scale datasets. Emphasis is also placed on improving model interpretability and aligning predictions with human perception.