Pure shift NMR spectroscopy enables the robust probing on molecular structure and dynamics, benefiting from great resolution enhancements. Despite extensive application landscapes in various branches of chemistry, the long experimental times induced by the additional time dimension generally hinder its further developments and practical deployments, especially for multi‐dimensional pure shift NMR. Herein, this study proposes and implements the fast, reliable, and robust reconstruction for accelerated pure shift NMR spectroscopy with lightweight attention‐assisted deep neural network. This deep learning protocol allows one to regain high‐resolution signals and suppress undersampling artifacts, as well as furnish high‐fidelity signal intensities along with the accelerated pure shift acquisition, benefitting from the introduction of the attention mechanism to highlight the spectral feature and information of interest. Extensive results of simulated and experimental NMR data demonstrate that this attention‐assisted deep learning protocol enables the effective recovery of weak signals that are almost drown in the serious undersampling artifacts, and the distinction and recognition of close chemical shifts even though using merely 5.4% data, highlighting its huge potentials on fast pure shift NMR spectroscopy. As a result, this study affords a promising paradigm for the AI‐assisted NMR protocols toward broader applications in chemistry, biology, materials, and life sciences, and among others.

Accurate 3D medical image segmentation is crucial for diagnosis and treatment. Diffusion models demonstrate promising performance in medical image segmentation tasks due to the progressive nature of the generation process and the explicit modeling of data distributions. However, the weak guidance of conditional information and insufficient feature extraction in diffusion models lead to the loss of fine-grained features and structural consistency in the segmentation results, thereby affecting the accuracy of medical image segmentation. To address this challenge, we propose a Mamba-Enhanced Diffusion Model for 3D Medical Image Segmentation. We extract multilevel semantic features from the original images using an encoder and tightly integrate them with the denoising process of the diffusion model through a Semantic Hierarchical Embedding (SHE) mechanism, to capture the intricate relationship between the noisy label and image data. Meanwhile, we design a Global-Slice Perception Mamba (GSPM) layer, which integrates multi-dimensional perception mechanisms to endow the model with comprehensive spatial reasoning and feature extraction capabilities. Experimental results show that our proposed MambaDiff achieves more competitive performance compared to prior arts with substantially fewer parameters on four public medical image segmentation datasets including BraTS 2021, BraTS 2024, LiTS and MSD Hippocampus. The source code of our method is available at https://github.com/yuliu316316/MambaDiff.

As a perfect complement to conventional NMR that aims for chemical structure elucidation, Laplace NMR constitutes a powerful technique to study spin relaxation and diffusion, revealing information on molecular motions and spin interactions. Different from conventional NMR adopting Fourier transform to deal with the acquired data, Laplace NMR relies on specially designed signal processing and reconstruction algorithms resembling the inverse Laplace transform, and it generally faces severe challenges in cases where high spectral resolution and high spectral dimensionality are required. Herein, based on the tensor technique for high-dimensional problems and the sparsity assumption, we propose a general method for high-resolution reconstruction of multidimensional Laplace NMR data. We show that the proposed method can reconstruct multidimensional Laplace NMR spectra in a high-resolution manner for exponentially decaying relaxation and diffusion data acquired by commercial NMR instruments. Therefore, it would broaden the scope of multidimensional Laplace NMR applications.

Nuclear magnetic resonance (NMR) spectroscopy is widely used across chemistry, applied physics, life sciences, and related disciplines. As NMR studies grow in complexity, artificial intelligence (AI) has emerged as a transformative tool to improve NMR data acquisition, processing, and analysis, fundamentally reshaping conventional NMR workflows. This review provides a comprehensive overview of recent advances in AI-enabled NMR reconstruction, tracing its methodological evolution from early artificial neural networks and evolutionary algorithms to contemporary deep learning (DL) frameworks. Main applications are examined in detail, including sparse reconstruction, noise filtering and artifact suppression, Laplace NMR inversion, pure shift NMR, chemical exchange saturation transfer NMR, RF pulse generation and pulse sequence design, and nanoscale NMR, among others. For each of these applications, AI methodologies, design choices, key innovations, and publicly available data repositories are highlighted. Moreover, we also summarize and compare the technical implementations and quality assessment behind these applications. Finally, we discuss current challenges, including trade-off between signal preservation and artifact suppression, limited model generalizability to unseen data, the absence of online and uniform quality assessment metrics, and the scarcity of high-quality experimental datasets, and outline future directions encompassing advanced network architectures and training strategies, the development of foundation models for NMR reconstruction, uncertainty-aware modeling and quality assessment benchmarking platforms, and the establishment of open-source datasets. Collectively, the integration of AI addresses long-standing limitations in NMR spectroscopy and improves the quality of NMR spectra, enabling automated analysis of experimental data and enhancing subsequent spectral interpretation, thus providing the stronger support for scientific research and practical applications.