# US 11,906,965 B2 · System and method for conscious machines

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,906,965 B2
- Document ID: US11906965B2
- Status: Granted
- Kind: B2
- Application number: 17/447,454
- Filed: 2021-09-13
- Published: 2022-01-06
- Granted: 2024-02-20
- Earliest claimed: 2018-02-22
- Applicants: Alan M. Kadin
- Inventors: Alan M. Kadin
- Web page: https://patentgrep.com/patent/US11906965B2

## CPC

- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G05D 1/00: Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots (drive control systems specially adapted for autonomous road vehicles B60W60/00)
- G05D 1/22: Command input arrangements
- G05D 1/228: Command input arrangements located on-board unmanned vehicles
- G06F 18/2178: {based on feedback of a supervisor}
- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 3/092: Reinforcement learning
- G06N 5/042: {Backward inferencing}
- G06N 5/043: Distributed expert systems; Blackboards
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G05D: SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Consciousness is widely considered to be a mysterious and uniquely human trait, which cannot be achieved artificially. On the contrary, a system and method are disclosed for a computational machine that can recognize itself and other agents in a dynamic environment, in a way that seems quite similar to biological consciousness in humans and animals. The machine comprises an artificial neural network configured to identify correlated temporal patterns and attribute causality and agency. The machine is further configured to construct a virtual reality environment of agents and objects based on sensor inputs, to create a coherent narrative, and to select future actions to pursue goals. Such a machine may have application to enhanced decision-making in autonomous vehicles, robotic agents, and intelligent digital assistants.

## Claims

### Claim 1

A computational system configured to interact with a physical environment, comprising:
at least one input device configured to receive information relating to a robot self agent, other agents, and objects in the physical environment;
at least one output device configured to produce an output to change a relationship of at least one of the robot self agent, other agents, and non-agent objects in the physical environment; and
at least one automated processor, configured to:
establish at least one goal based on at least a set of rules for interacting between the robot self agent and the other agents and objects in the physical environment;
periodically update a simplified dynamical predictive model of the physical environment which provides a coherent narrative of the robot self agent interacting with the other agents and the objects in the physical environment, based on at least the received information, and store a series of the simplified dynamical predictive models in a memory;
the simplified dynamical predictive model being configured to predict a plurality of alternative future states of the robot self agent, the other agents, and the objects within the physical environment; and
decide between prospective alternate actions of the robot self agent, based on at least the at least one goal, the stored simplified dynamical predictive models of the physical environment, and the plurality of alternative future states.

### Claim 2 (depends on 1)

The computational system according to claim 1, wherein the output is dependent on the decision between the prospective alternate actions.

### Claim 3 (depends on 1)

The computational system according to claim 1, wherein the output is dependent on the periodically updated simplified dynamical predictive models and independent of the decision between prospective alternate actions of the robot self agent.

### Claim 4 (depends on 1)

The computational system according to claim 1, wherein the at least one automated processor is further configured to process the received environmental information and recognize changes in the robot self agent, the other agents, and the objects in the physical environment, wherein the simplified dynamical predictive model is responsive to the recognized change.

### Claim 5 (depends on 1)

The computational system according to claim 1, wherein the at least one automated processor is comprises a neural network having a plurality of hidden layers between input and output.

### Claim 6 (depends on 1)

The computational system according to claim 1, further comprising a clock configured to define a period for the periodic updates of the representation.

### Claim 7 (depends on 6)

The computational system according to claim 6, wherein the at least one automated processor is further configured to retrieve simplified dynamical predictive models of the physical environment from the memory by time-stamp and by a content of the simplified dynamical predictive models.

### Claim 8 (depends on 1)

The computational system according to claim 1, wherein the robot self agent further comprises a navigational process for navigating the robot self agent in the physical environment.

### Claim 9 (depends on 8)

The computational system according to claim 8, where the robot self agent comprises at least one of an autonomous vehicle and a robot.

### Claim 10 (depends on 1)

The computational system according to claim 1, wherein the other agents are autonomous navigation mobile devices.

### Claim 11 (depends on 1)

The computational system according to claim 1, further comprising a virtual environment in which the robot self agent can navigate, for training of the computational system for operation in the physical environment.

### Claim 12 (depends on 1)

The computational system according to claim 1, wherein the at least one automated processor is further configured to communicate with the other agents in the physical environment comprising human agents and computational agents.

### Claim 13 (depends on 1)

The computational system according to claim 1, further comprising an artificial neural network having a plurality of layers comprising at least one hidden layer for the deciding between prospective alternate actions of the robot self agent.

### Claim 14 (depends on 13)

The computational system according to claim 13, wherein the artificial neural network is implemented using at least one of the group consisting of at least one of a tensor processing unit, a graphics processing unit, a field programmable gate array, and a neuromorphic chip.

### Claim 15

A computational method for interacting with a physical environment, comprising:
receiving information relating to a robotic self agent, other agents, and objects in the physical environment;
establishing at least one goal based on at least a set of rules for interacting between the robotic self agent and the other agents and objects in the physical environment;
periodically updating a simplified dynamical predictive model of the physical environment which provides a coherent narrative of the robotic self agent interacting with the other agents and the objects in the physical environment, based on at least the received information, and storing a series of the simplified dynamical predictive models with at least one automated processor;
predicting a plurality of alternative future states of the robotic self agent, the other agents, and the objects within the physical environment with a causal model with the at least one automated processor;
deciding between prospective alternate actions of the robotic self agent with the at least one automated processor, based on at least the at least one goal, the stored simplified dynamical predictive models, and the plurality of alternative future states; and
controlling the robotic self agent to interact with the other agents and the objects in the physical environment, dependent on at least the prospective alternate actions of the robotic self agent, by producing an output for changing a relationship of at least one of the robotic self agent, the other agents, and the objects in the physical environment.

### Claim 16 (depends on 15)

The computational method according to claim 15, wherein the output for changing the relationship is dependent on the decision between the prospective alternate actions.

### Claim 17 (depends on 15)

The computational method according to claim 15, wherein the output for changing the relationship is dependent on the simplified dynamical predictive models and independent of the deciding between prospective alternate actions of the robotic self agent.

### Claim 18 (depends on 15)

The computational method according to claim 15, further comprising processing the received environmental information to recognize changes in the robotic self agent, the other agents, and the objects in the physical environment, wherein the simplified dynamical predictive model is responsive to the recognized change.

### Claim 19 (depends on 15)

The computational method according to claim 15, wherein the at least one processor comprises a neural network having a plurality of hidden layers between input and output.

### Claim 20

A method for controlling interactions of a robotic agent within a physical environment, comprising:
defining a set of rules for interacting between the robotic agent and the physical environment;
recognizing active other agents in the physical environment, objects in the physical environment, and the robotic agent in the physical environment;
recognizing changes in the physical environment using pattern recognition based on an environmental status input;
sequentially generating and storing a series of simplified dynamical predictive models of the physical environment comprising other agents, the robotic agent, and the objects;
defining at least one goal for the robotic agent based on at least the set of rules;
making alternative predictions of future states of the other agents, the robotic agent, and objects within the physical environment using a simplified dynamical predictive model, based on alternate different actions by the robotic agent and past representations stored in the memory with at least one automated processor;
generating decisions based on the goals and the alternate predictions with the at least one automated processor;
controlling the robotic agent, the agents in the physical environment, and the objects in the physical environment, dependent on the generated decisions, to change a relationship; and
updating the simplified dynamical predictive model to present a dynamic narrative of the robotic agent within the physical environment.