Institute · 01
The Institute for AI & Machine Learning advances the science of intelligent systems — from deep learning theory to safe, aligned AI deployable at civilizational scale. Our three laboratories cover the full spectrum from foundational algorithms to real-world deployment.
Laboratories
Each lab operates independently with dedicated staff, shared institute infrastructure, and its own research agenda.
Theoretical and empirical research on deep neural architectures — transformers, diffusion models, sparse networks, and the mathematics of generalization in high dimensions.
Ensuring that powerful AI systems reliably pursue intended goals. Research spans RLHF, constitutional methods, oversight scalability, and formal specification of human values.
Unified representations across language, vision, audio, and structured data. Research on cross-modal grounding, zero-shot transfer, and embodied language understanding.
Selected Publications