FAQs
Q: What is Cyquential AI studios, and what is Cyquential AI?
A: Cyquential AI Studios is a company based around games and artificial intellegence. Cyquential AI is a new specific type of AI that is made by the Cyquential Studios Company
Cyquential AI FAQ
Q: What seperates Cyquential AI from tradditional AI in machine learning and deep learning?
A: In machine learning, typically learning is done through an artificial neural network, where each neuron represents a number, and the number of neurons is fixed. In contrast, a Cyquential neural network uses first order propositional logic (FOL) to represent each neuron, and continuously builds new neurons as it learns and gains more experience. This is different from regular artificial neural networks as the number of neurons is fixed. It is similar in concept to a Markov Logic Network, and Inductive Logic Programming
Q: How do Cyquential AI learn? A: Cyquential AI use a new learning hypothesis called Cyquential Hypothesis (hence the name Cyquential AI), which combines associative learning (from physology) with hebbian learning (from neuroscience). It uses this learning theory to continously create and build more neurons off of existing ones.
Q: what can Cyquential AI do? A: Currently, the Cyquential AI can decently recognize hand written digits (from the mnist dataset), posses basic object understanding and can grasp basic shapes. It is also a reinforcement learning agent that performs semi decently on super mario bros, pong, and more. (edited)
Q: What are the uses of Cyquential AI?
A: Right now, the biggest and most direct application of Cyquential AI is in universal recognition systems and transfer learning. Cyquential AI don't require any specific datasets and don't need a teaching signal. To learn, they simply roam around the enviornment and gain information. When learning new concepts, all the previous concepts are used, and this specific technique makes transfer learning and univeral object recognition systems possible. Cyquential AI only took about 12 samples to learn the MNIST dataset, but in addition, it could have also learned any other character, such as chinnese characters or shape drawings or the english alphabet. Showing any of these to Cyquential AI costs very few training examples, and Cyquential AI learns them almost instantly, while still retaining all of it's previous information, such as the previous examples it was shown
Q: How is the Cyquential AI Implemented?
A: Currently, the Cyquential AI is programmed in python (v.3.7.4), using Webots (an open source robotics simulator) as it's envrionment to explore. In addition, the project uses OpenCV for general purpose preprocessing and graphics, numpy for various mathmatical operations, and networkx + matplotlib for visualizations. There is also a c++ library being worked on.
Q: Can I use the Cyquential AI today?: A: Currently, the Cyquential project is private. However, the goal is that it eventually becomes public, after enough work is done to iron the architecture out.
Q: Is there any information on Cyquential Hypothesis?
A: The majority of the contents of Cyquential Hypothesis are in a book that is still being worked on that is currently 230+ pages. It is planned to be fully released aroud next year. In addition, I will be presenting at various AI conferences around the USA, so more information will be avalible there (dates will be anounced soon).
Psychophrenesis FAQ
Q: What is Psychophrensis?:
A: The game is also still in the works! It is a first/third person survival shooter made with Unreal Engine 5.
Q: Could I help work on Cyquential AI or the game? A: Yes! I am always looking for new people to come help out. If you would like to join the staff of Cyquential AI Studios, please message me for more information. I am avalible through discord dm's, or through email at mallory.115@osu.edu Work is both paid and unpaid, depending on the feature.