First Principles AI
We are offering a patented (USPTO) class of solutions to sequential data modeling problems, including sequence-to-sequence modeling and reinforcement learning.
At the heart of our solution is a computational mechanism that allows for scalable solutions to sequential data modeling problem without the need for expensive neural solutions.
Sequential Data Modeling Problems
Sequence-to-Sequence Modeling
Accurate sequence-to-sequence modeling orders-of-magnitude faster than neural approaches
Sequence-to-Sequence Modeling with State Magnitudes
Model a system comprised of ordered sequences of states where each state also has an associated measurable magnitude
The Treasure Hunt Problem - Decision Processes with Deep History
Model a decision process system where the full state and action history is relevant to the next action
Novel, Patented Solutions in Sequence-to-Sequence Modeling and Reinforcement Learning
First Principles AI offers leases to intellectual property rights and software for sequential data modeling solutions. Designed and patented by physicist and data science executive Jason Nett, Ph.D., our solution can model sequence-to-sequence modeling problems at least four orders-of-magnitude more efficiently than neural-based approaches and a wider class of reinforcement learning problems where the order of rewards achieved is intrinsic to the environment being modeled.
Jason Nett earned a Ph.D. in High Energy Physics in 2010 from the University of Wisconsin-Madison with a study contributing to the Higgs Boson search at the Fermi National Accelerator Laboratory. After an additional three years in academia as a postdoctoral research associate studying supersymmetry, he moved to the private sector in 2013 where he has worked in the data and AI space building tech startups.
Accurate and Efficient
Our patented solutions push the limits of imagination for accuracy and efficiency in sequential data modeling.
Solutions to New Problems
We offer solutions to problems that have largely not been addressed by industry or academia.
Unique and Novel
Our solutions are completely novel and designed from first principles.