The recent advancements in the field of audio and music information retrieval are significantly pushing the boundaries of what is possible with machine learning and deep learning techniques. A notable trend is the increasing use of large language models (LLMs) and transformer-based architectures, which are being leveraged for tasks such as music genre classification, emotion recognition, and multilingual music information retrieval. These models are demonstrating superior performance over traditional deep learning architectures, particularly in zero-shot scenarios and cross-dataset label alignment. Additionally, there is a growing focus on real-time processing and the integration of multimodal data, such as in event-centric video retrieval, where systems are required to synthesize information from visual, audio, and textual sources. The field is also seeing innovative applications in conservation efforts, with automated systems for detecting and classifying animal calls, such as elephant vocalizations, which can provide valuable insights for environmental management. Furthermore, computational analysis of traditional music forms, like Pansori singing, is being explored to uncover unique audio characteristics and facilitate documentation and education. Overall, the integration of advanced machine learning techniques with multimodal data is paving the way for more robust and inclusive systems in audio and music information retrieval.