LabOS: The AI-XR Co-Scientist That Sees and Works With Humans
Cong, L.; Smerkous, D.; Wang, X.; Yin, D.; Zhang, Z.; Jin, R.; Wang, Y.; Gerasimiuk, M.; Dinesh, R. K.; Smerkous, A.; Shi, L.; Zheng, J.; Lam, I.; Wu, X.; Liu, S.; Li, P.; Zhu, Y.; Zhao, N.; Parakh, M.; Serrao, S.; Mohammad, I. A.; Chen, C.-Y.; Xie, X.; Chen, T.; Weinstein, D.; Barbone, G.; Caglar, B.; Sunwoo, J. B.; Li, F.; Deng, J.; Wu, J. C.; Wu, S.; Wang, M.
10.1101/2025.10.16.679418 bioRxivShow abstract
Modern science advances fastest when thought meets action. LabOS represents the first AI co-scientist that unites computational reasoning with physical experimentation through multimodal perception, self-evolving agents, and XR-enabled, embodied human-AI collaboration. To empower agentic AI with embodied intelligence, we built LabSuperVision, the first vision-language-model (VLM) benchmark spanning biomedical and materials science laboratories, designed to evaluate scientific perception and reasoning. This endowed LabOS with visual reasoning capabilities that surpass those of general-purpose models on laboratory-specific tasks. By connecting multi-model AI agents, smart glasses, and human-AI-robot collaboration, LabOS allows AI to see what scientists see, understand experimental context, and participate in real-time execution through immersive XR interaction and robotic autonomation. Across applications-from cancer immunotherapy target discovery to cell engineering-LabOS shows that AI can move beyond computational design to participation, turning the laboratory into an intelligent, collaborative environment where human and machine discovery evolve together.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Automatic mapping of multiplexed social receptive fields by deep learning and GPU-accelerated 3D videography 94%
- Large-scale capture of hidden fluorescent labels for training generalizable markerless motion capture models 94%
- OmicVerse: A single pipeline for exploring the entire transcriptome universe 93%
Similar papers in this journal
- CRISPR-GPT for Agentic Automation of Gene Editing Experiments 92%
- Identifying perturbations that boost T-cell infiltration into tumours via counterfactual learning of their spatial proteomic profiles 92%
- Combinatorial prediction of therapeutic perturbations using causally-inspired neural networks 90%
Similar papers in this journal
- scvi-tools: a library for deep probabilistic analysis of single-cell omics data 93%
- Whole-cell segmentation of tissue images with human-level performance using large-scale data annotation and deep learning 93%
- scJoint: transfer learning for data integration of atlas-scale single-cell RNA-seq and ATAC-seq 92%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.