Signal Reconstruction

Signal reconstruction aims to recover a complete signal from incomplete or noisy measurements, a crucial task across diverse scientific fields. Current research emphasizes developing advanced algorithms and model architectures, such as invertible neural networks, unrolled neural networks, and masked autoencoders, to improve reconstruction accuracy and efficiency, often incorporating prior knowledge about signal structure or employing joint sampling and reconstruction strategies. These advancements have significant implications for applications ranging from biomedical signal processing (e.g., EEG and MRI reconstruction) to communication systems (e.g., image and RF signal recovery) and environmental monitoring (e.g., air pollution data imputation).