We are developing a neuromorphic electronics assembly that will provide some of the useful attributes and capabilities of human conscious behavior to machines and devices. Artificial intelligence, and computer science in general, is increasingly evolving along the lines of massively parallel and interconnected processing. A means is needed to resolve conflicts and collisions in such systems. To quote David Eagleman1, consciousness “plays a vital role in settling conflicts within the brain.” “Consciousness is the system that has this unique vantage point, one that no other subsystem of the brain has. And for this reason, it can play the role of arbiter of the billions of interacting elements, subsystems, and burnt-in processes.” We have discovered the means to provide this executive oversight to machines and devices.
Conscious decision making is taken for granted by humans along with the sense of self as the decision making agent. Until recently, little was known of the neuronal correlates of conscious decision making to guide the design of artificial versions. Finally, after decades of pursuit, neuroscience has provided a definite clue.2 A distributed assembly of neurons with connections to all of the sensors and motor centers acting as a single unit is the architectural foundation for the machine analog for human conscious decision making called Hyper Aware Logic, or for short. It is difficult to imagine navigating the real world unconsciously, like a sleep walker, except that is what robots and other autonomous machines are required to do. is a platform that can be retrofitted to existing machines to, basically, wake them up. With , the inherent latency in laborious unconscious logical processes can be overcome. responds real-time to the sensed internal and external environments to correct, enhance, or just continue a plan of action that was developed and initiated unconsciously.
The defining characteristic of this new paradigm is the dichotomy between a conscious, decision making brain that contains the learned understanding of good and bad outcomes, and an unconscious site of true intelligence that includes deep learning, planning, vocabulary, and language. The conscious brain has little to do with intelligence, but a great deal to do with value judgements. An eagle’s conscious brain is at least as advanced as our own, but its intelligence and communication skills are meager. Consciousness does not equate to intelligence but without consciousness, intelligence is crippled.
In this paradigm, consciousness is enabled by unconscious understanding; i.e. one cannot see what one cannot comprehend. The reason for this is easy to grasp, since otherwise a conscious decision maker would have no basis for any decision based upon sensory input. A well-known example of this phenomenon is blind sight, where a lesion in the visual processing pathway deprives the victim of conscious sight in the related field of view even though reflexive responses to threats in that field still occur. Decisions to duck a punch are always unconscious, anyway. Understanding is critical to the conscious brain’s grouping of sensed things, events, and activities for the downstream neural processing preparatory to decision making.
In animal brains it is becoming generally accepted that the thalamus nuclei that route nearly all sensory outputs to their various cortices comprise the likely seat of consciousness. The coupling of consciousness to the decision making process is suggested by the large volume of excitatory and inhibitory signals from the thalamus to the motor centers. This new paradigm for conscious decision making in machines and devices is built upon a model of the animal thalamus.
The conceptual breakthrough is the answer to the question: “How can we make the self-aware agent for decision making?” The answer is by distributing the sensor outputs to each of the decision network logic units in such a way as to linearly influence its behavior and to have the entire network address one decision at a time acting as a single entity.
is the piece that has always been missing from the decades long pursuit of Artificial Intelligence (AI). It provides personhood, a sense of self, a decision making first-person singular directly connected to each sensor in the body. With , your “smart” phone will morph into Radar O’Reilly, anticipating your needs and accessing all of the means to provide them while remaining aware of everything you are aware of and much more.
Even with all of the investment thus far in AI, the capabilities resident in available computers do not approach that of humans. The performance of an eagle is more reachable today than that of a human, but even eagles are much more capable than any robot. This is because eagles are conscious and self-aware. We will demonstrate initially at the Eagle level using our ALERT® product to emulate the unconscious brain capabilities. Going forward, we will turn our smart phones into Radar O’Reilly using the cloud as our unconscious brain.