We are developing and testing new model architectures & network primitives to achieve massively multimodal, “Generalist” systems capable of doing useful things across images, natural language, and the physical and digital world.
Large scale, generalist architectures capable of tasks across Images, Text, and Control Domains.
We're building new kinds of action models capable of using digital tools.
We’re designing new kinds of Agent Architectures, pulling from ideas in Cognitive Science and Reinforcement Learning to equip AI Agents with memory, modelling, and other key capabilities, including the use of digital tools and software.
We're building new kinds of action models capable of using digital tools.
Survey and Perspective Paper on Opportunities in AI Agents
Robots are embodied intelligent systems with almost limitless real world applications, but a large accompanying set of challenges. Instantiation of intelligence in the real world brings with it noise, disturbances, actuation and performance limitations, incomplete information, and more. We’re developing theory and tooling to understand robotic cognition and guide performance under these constraints.
Enabling multi-agent systems to cooperate on tasks, with an emphasis on real time adaptation to out-of-distribution changes.
Investigating higher order thinking in robotic agents, and how perception and actuation affect cognition.
We’re exploring the application of foundation class models to biological problems, including drug discovery, feature detection, protein mapping, and more. Currently in an exploratory capacity, stay tuned for project details. Follow our progress on Discord