News
Local pseudopotentials accelerate neural network wavefunction simulations - Nature Computational Science
1+ hour, 29+ min ago (184+ words) Nature Computational Science (2026) Cite this article Hermann, J. et al. Ab initio quantum chemistry with neural-network wavefunctions. Nat. Rev. Chem. 7, 692–709 (2023). This review presents the early development of neural network wavefunction simulations and their emergence as a practical approach to ab initio…...
Which ‘AI scientist’ suits your lab? A guide for the perplexed
3+ day, 16+ hour ago (338+ words) General purpose AI tools for science, such as Claude Science, promise to accelerate research. But which is right for you? General purpose AI tools for science, such as Claude Science, promise to accelerate research. But which is right for you?...
Health system learning enables generalist neuroimaging models | Nature Medicine
3+ day, 19+ hour ago (1662+ words) Nature Medicine (2026) Cite this article In summary, NeuroVFM demonstrates that generalist AI models can be built from the health system itself. Learning directly from health system data yields representations that translate to diagnosis, triage and report generation. We provide a…...
Judicious use of LLMs could speed up progress in the social sciences
5+ day, 12+ hour ago (46+ words) Large language models could advance experiments in the social sciences, but fundamental questions remain about how they should be used to support research. Large language models could advance experiments in the social sciences, but fundamental questions remain about how they…...
Universal cell embedding provides a foundation model for cell biology
5+ day, 13+ hour ago (1688+ words) Nature (2026) Cite this article UCE is trained in a fully self-supervised manner, without using any cell type or dataset-based annotations. While incorporating cell type labels into the model design or training could seem advantageous, it is preferable for the organization…...
msBayesImpute as a versatile framework for addressing missing values in biomedical mass spectrometry proteomics data | Communications Chemistry
6+ day, 13+ hour ago (1473+ words) To further evaluate robustness under extreme missingness, we generated datasets with 50 samples and missing rates ranging from 10 to 90%. Imputation error (RMSD, root mean squared deviation) increased sharply when the missing rate rose from 70 to 80%, indicating that msBayesImpute performs robustly up…...
AI decodes protein–ligand binding - Nature Chemical Biology
1+ week, 19+ hour ago (130+ words) Nature Chemical Biology (2026) Cite this article Wang, L. et al. J. Am. Chem. Soc. 137, 2695–2703 (2015). Shen, C. et al. Chem. Sci. 14, 8129–8146 (2023). Mastropietro, A., Pasculli, G. & Bajorath, J. Nat. Mach. Intell. 5, 1427–1436 (2023). Yu, J. et al. Nat. Comput. Sci. 3, 860–872 (2023). Landrum, G. A. & Riniker, S. J....
Collection policies | From Theory to Application: Advances in Multi‑Agent Systems/Frameworks
1+ week, 3+ day ago (185+ words) This Collection in npj Artificial Intelligence seeks to showcase cutting‑edge research that bridges algorithmic innovation with practical deployment. We invite contributions that highlight both the conceptual depth of MAS frameworks and their applied impact in diverse environments. This Collection…...
About the Guest Editors | From Theory to Application: Advances in Multi‑Agent Systems/Frameworks
1+ week, 3+ day ago (137+ words) This Collection in npj Artificial Intelligence seeks to showcase cutting‑edge research that bridges algorithmic innovation with practical deployment. We invite contributions that highlight both the conceptual depth of MAS frameworks and their applied impact in diverse environments. This Collection…...
Why high scores do not mean application readiness for health AI - Nature Medicine
1+ week, 4+ day ago (160+ words) Nature Medicine (2026) Cite this article Version of record: 02 July 2026 Singhal, K. et al. Large language models encode clinical knowledge. Nature 620, 172–180 (2023). A landmark study reporting strong LLM performance on clinician benchmarks. Results of this kind motivated optimism about clinical readiness, which our…...