The field of audio analysis and biosignal processing is rapidly evolving, with a focus on developing innovative methods for improving the accuracy and efficiency of various applications. Recent studies have explored the use of machine learning techniques, such as deep learning and neural networks, to analyze audio signals and biosignals for disease detection, pain assessment, and environmental sensing. Notably, researchers have proposed novel approaches, including the integration of frequency selection and attention mechanisms, hyperbolic embeddings, and combolutional neural networks, to enhance the performance of audio analysis tasks. Additionally, the development of lightweight embedding models, such as Tiny-BioMoE, has shown promise in biosignal analysis. The use of Wi-Fi signals for passive sensing and respiration monitoring has also been investigated, demonstrating the potential for low-cost and non-intrusive monitoring solutions. Overall, these advancements have the potential to significantly impact various fields, including healthcare and environmental monitoring.

Noteworthy papers include: Improving Deep Learning-based Respiratory Sound Analysis with Frequency Selection and Attention Mechanism, which proposes a compact CNN-Temporal Self-Attention network for efficient respiratory sound analysis. Comvolutional Neural Networks, which introduces a novel combolutional layer for extracting harmonic features in audio signals. Tiny-BioMoE, which presents a lightweight pretrained embedding model for biosignal analysis. WiRM, which demonstrates a two-staged approach to contactless respiration monitoring using conjugate multiple channel state information and fast iterative filtering in Wi-Fi systems.