Author: Albert Su
Published: Jan 17, 2019

Artificial intelligence (AI) is rooted in the neurosciences where fundamental questions of neural networks, a subset in computer science, emerged from the study of the biological behavior of neurons. This neural network similar to a brains’ network of unique protein synapsis connected to consolidated short-term and long-term memories is distributed. Distributed from the hippocampus located at the center of the brain where short-term memory is processed is held there for several days and later transferred to various prefrontal cortex regions of the brain for long-term storage. In order to find an answer to how the brain distributes intelligence, transferring memories from one region to another region, a model of a distributed AI blockchain was discovered.

What is a distributed AI blockchain? A distributed AI (DAI) is a de-centralized AI multi-agent system where large-scale distributed nodes process in parallel different AI problems (pattern recognition, imagery, sound). Similarly, in the brain various regions process different functions such as speech, visual, and auditory. Likewise different nodes in a network process various functions.

What is a blockchain? A blockchain is a growing list of records where blocks are linked together using hashes in cryptography. The reason for the cryptographic hashes is to establish the relationship to the previous block which is timestamped contained within the hash. This relationship is represented by a chain of blocks and recorded in a distributed ledger which is managed by a peer-to-peer network. To compare and contrast to a biological model, this relationship between short-term and long-term memory is similar to a distributed blockchain ledger.

What are the applications for distributed AI blockchain? While traditional models of AI have had been largely centralized and large amounts of CPU computing power needed to processes the AI and blockchain, a better solution was needed. A de-centralized model that processes AI with multi-agent nodes over a distributed network solution was sought. For example, in one application for autonomous vehicles, this intelligence can be distributed from various portions of the vehicle such as fuel consumption, speed, and environmental conditions to calculate the proper speed and direction of the vehicle. In other AI applications such as visual odometry, auditory recognition, and sensory perception is to be explored in robotics, drones, and other autonomous systems.