Knowm Inc. is an emerging start-up company with the goal of integrating Nature’s transistor via memristors on silicon CMOS chips and consequently creating the world’s most powerful and energy efficient learning processors.
For more than a decade we have been developing the theory of AHaH Computing as described in great detail in a PLoS One paper from early 2014. With the Knowm API, a machine learning library built on top of a Thermodynamic-RAM emulator, we have shown state-of-the-art benchmark performance on several popular classification datasets as well as capabilities in robotic control, anomaly detection, signal prediction, and combinatorial optimization. Our vision is to combine the technological groundwork already put in place with the Knowm Development Community (KDC) for the creation of a self-optimizing global innovation engine where individuals and companies work together to develop modular reusable building blocks that can be rapidly deployed on real-world machine learning problems. A possible result of this collaborative effort is the creation of efficient general artificial intelligence, which we define as combined neuromemristive hardware-software systems capable of processing arbitrary data streams and taking actions in the real or virtual world to optimize arbitrary objective functions. Collaborating with top experts at Boise State University and Rochester Institute of Technology, Knowm is now transitioning into production mode, supported by a large intellectual property portfolio. With our first products such as commercially-available memristors, specially-designed hardware development platforms, the Knowm API and training tools available to collaborators, we are now taking the first steps towards our end goal of creating the world’s most powerful and energy efficient learning processors. We invite you to follow our progress, let your voice be heard and collaborate directly with us for profit and fun!