Harrisburg University-backed Thought Network believes that smart data can be processed a lot cheaper and faster by embedding every piece of data with artificial intelligence, according to a company statement announcing an ICO.
The start-up is challenging the idea that in order for data to become useful it needs an application, whereas a “hybrid data agent” is able to act on its own — understanding where it came from, where it’s supposed to go, and what it has to do.
Thought’s website describes the architecture as a “Universal Turing Machine”; a lofty claim indeed.
“Traditionally, data just sits around until it is sent to applications that are able to process that information, but Thought makes data agile and able to act on its own without any external help,” founder and CEO of Thought, Andrew Hacker said in a statement. Hacker is also a cybersecurity expert in residence at Harrisburg University.
Integrating blockchain technology with this concept creates an extra layer of security on the smallest level. Every piece of data is secured with cryptography, which means the owners of the data to dictate exactly who has access to it.
Solving the “black box” problem
One of the big challenges facing AI development is complexity, and designers of the neural networks admit that they are too complex to understand how the outputs were derived. This is often referred to as the “black box problem”.
Since the training of data and AI application are combined, AI training is simplified and the black box problem is solved by hybrid data that knows how it traveled through a neural network, according to Thought.
The smallest, and a critical, part of the network are “nuances”, which are comprised of multiple parts, including data slices and application code, according to Thought’s whitepaper.
Nuances are the vehicles for transporting, organizing, sending, analyzing, and tracking data across the network and ultimately the globe. They are “intelligent”, that is, they are encoded with AI capability that allows them to “understand” how to behave.
Based on this capability combined with their encoded logic, they execute intended behavior, typically in groups that have the capability to intelligently gather, swarm, or flow through the network in the most efficient way possible.
For example, a set of Nuances may be used to correlate two input streams of data and find the irregular patterns between the two streams. As such, the AI will understand that this path is the correct one for its lifecycle and route through the network automatically without needing to pass through gateway machines.
What do data that think wear?
The technology stack contains a “fabric layer”, which represents the protocol that facilitates the activity of the network. Nuances exist within an infinitely scalable network that allows each bit of artificial intelligence to interact forming an intelligent system.
Nuances working within Fabric exist as electronic neurons within a globally distributed brain. Systems that were once incompatible become integrated, and disparate processes and infrastructure become unified, according to the company.
“The upcoming release of this new technology, Thought Blockchain, and its applications for businesses and consumers in artificial intelligence and analytics represents the bleeding edge of innovation,” said Eric Darr, president of Harrisburg University in a statement.