Towards Scientific Artificial Intelligence
Improve human life through AI-driven scientific discovery
Our vision is that generative AI should be used to build progressively more complex scientific models that will provide explanation of the world, generate interesting testable hypotheses that will help improve and validate scientific models, and unleash fascinating scientific breakthroughs that will improve the life of the entire civilization on Earth and beyond
Find out about our startups:
TuringDatU?, a prototype of a human – model interface for reinforcement learning with gamified human feedback. This solution will help integrate human scientific knowledge into SciAI models
Quazality MoneyFlow, an example of a scientific causal model, representing a statistically-derived hypothesis of cash flows within the stock market, both on the industry-group and the individual-stock level. This solution demonstrates the possibility of deriving and validating the causal structure of abstract concepts in a SciAI model