A computational method called scSurv, developed by researchers at Institute of Science Tokyo, links individual cells to patient outcomes using widely available bulk RNA sequencing data. The approach uses single-cell reference datasets together with patient survival data to infer the contributions of individual cells within complex tissues. The model identified cell populations associated with survival across several cancers, offering a way to uncover disease-driving cells and support the development of more targeted treatment strategies.

Seamless integration between electronics and the human body is the goal, and Ga-LMs are key to this transformation. Essential properties,such as, fluidity, conductivity, and biocompatibility, enable Ga-LMs to form stretchable, self-healing circuits, paving the way for advanced wearables, soft robotics, and medical implants that promise to redefine human-machine interaction.

Neuroimaging analysis in brain disorders faces a persistent challenge: brain signals are complex and high-dimensional, while high-quality labeled datasets remain limited. This review article systematically examines how self-supervised learning can help address that gap by learning meaningful representations directly from unlabeled neuroimaging data. It covers major methodological families, including contrastive, generative, and hybrid generative-contrastive approaches, and discusses their applications in functional MRI, EEG, and multimodal brain network analysis.

The review argues that self-supervised learning offers more than annotation efficiency. It may enable more transferable and clinically useful representations for disease screening, diagnosis, and prognosis across heterogeneous datasets and disorders. At the same time, interpretability, data heterogeneity, missing modalities, and clinical validation remain major barriers. Future work will likely focus on stronger multimodal fusion, better cross-site generalization, and more clinically adaptable model design.

A research team from Uppsala University has developed a machine-learning strategy that gives much faster access to molecular electrostatic potentials, a fundamental property that governs how molecules interact with ions, solvents, and each other. By training equivariant graph neural networks on dipole and quadrupole moments, the researchers showed that AI models can recover electrostatic features much more accurately than dipole-only approaches. The work opens a practical route for AI-guided screening of solvents and electrolytes for next-generation batteries and other energy devices.