What Is Synthetic Sentience?

Animal brains exhibit sentience in nature. They sense the environment, and generate output by controlling muscles, which provide for locomotion, manipulation of the environment, and communication with other animals. Sentience is the closed-loop awareness that arises in brains when both their sensory data and their own generated behavior feed back as inputs, making the brain aware of its own interaction with the environment.

Synthetic sentience is the same effect, produced in a synthetic brain that uses robotic hardware or a virtual agent program to interact with its environment. These synthetic brains are really models of designed and engineered brains operating in a high-fidelity, real-time simulation engine.

What Can Synthetic Sentience Do For Robots?

The closed-loop learning-from-experience model that Synthetic Sentience (SS) makes it possible to make robots that can navigate their world autonomously, in a way that rule-based programming cannot offer. Just as Machine Learning (ML) makes it possible to create application programs that recognize faces or categorize pictures, SS provides real-time behavior associated with any set of sensory inputs and control outputs, through experience obtained in actual operation.

For example, adding terrain navigation to any robotic vehicle supplements the robot's primary tasks with additional sophistication: adaptive gait appropriate for the environment, balance, and even autonomous goal seeking.

In warehouse environments, adding object seeking behavior to a stock management robot enables it to keep looking for an item when it is not found in the expected location, taking the human out of the loop.

In environments dangerous to human life, such as deep water operations, space, fires, chemical spills, biological contagion zones, or radiation hazard areas, robots can perform surveillance, rescue, or repairs with autonomous operation, when no remote human operator is possible.

What Can Synthetic Sentience Do For Virtual Agents?

As robots without physical bodies, virtual agents are simply application programs connected to networks such as the Internet, allowing them to perform work by monitoring data on the networks, and calling network APIs as conditions dictate. Synthetic Sentience provides a level of unscripted autonomy that makes these agents more capable and resilient.

Already existing are fixed monitor applications for health of financial markets, network health, physical premise intrusion, and toxic gas detection. But while many of these applications may use ML for recognition of alarm conditions and notify a human while possibly shutting down operations, they do not include the capability to go investigate the problem. Investigative behavior, which might involve navigating other related data sources or physical locations, could allow the root cause of a problem to be discovered, saving lives and resources.