The term Human-Level Artificial Intelligence was first introduced by Alan Turing back in 1950 when he proposed the Turing test.
It has the same meaning as these terms: Strong AI, True AI, and Artificial general intelligence. these are different names referring to the same technology.
Human level AI is a robot that can learn information, think, and act like a human adult, with college level intelligence.
In 2006 i published a book on human level artificial intelligence and shortly after that filed back to back patents on this technology. This video describes in detail how my invention works. It’s about 2 hours long and it is a comprehensive guide to human level A.I.
In this video we will cover these topics:
1. Robot doing and managing recursive tasks (all human tasks are recursive by nature). When the robot activates linear thoughts in his mind, it creates a self-created computer program to manage tasks. This self-created software is modifiable and adaptable. For instance, if the robot has his left arm chopped off, he can still make a sandwich. If the robot is playing a video game and the kitchen is on fire, he will run out the house.
This self-created software inside the robot’s mind manages recursive tasks by forming primarily 4 containers: 1. task container. 2. rules container. 3. planning container. and 4. identity container. The robot uses logic and reasoning to modify these containers to take action.
2. Robot learning knowledge or skills by reading books or going to school. Teachers in school teach the robot how to seek out knowledge (goals, rules, and procedures) through books or lectures.
3. Robot practicing to play a game. Teachers also teach the robot how to “teach itself” how to practice a game. For example, if the robot learns the rules of basketball, the next step is to use that knowledge and play the game.
If you think about it. If the robot can learn knowledge about a new skill and learn to practice a skill, by itself, without direct guidance from teachers or pre-defined functions, then he can essentially learn any human skill. He can learn sports, algebra, calculus, computer science, and discrete mathematics.
This robot learns information the long way. He needs to attend school, from kindergarten to college, to learn all knowledge.
4. Learning information in terms of a bootstrapping manner, whereby old information is used to learned new information and knowledge in the robot’s brain recursively builds on top of each other to form complex intelligence. For example, first, the robot learns basic grammar. Then he uses that knowledge to write a sentence. Next, he uses that knowledge to write a paragraph. Then, he uses that knowledge to write a letter. Finally, he uses that knowledge to write a book. When the robot is writing a book, he is repeatedly accessing previously learned skills, such as writing a paragraph, writing a sentence, and knowledge on grammar rules.
5. Data stored in the robot’s brain is global data, not local data. If using a neural network (aka deep learning), data is stored as local data. The robot’s brain has to store and access data and skills globally.
A few years back, a bunch of neuro-scientists did extensive research on how the human brain stores cat images, and they discovered that when a cat image is recognized by a human, his entire brain lights up. This basically means, cat images are stored in different areas of the brain. If the human brain is using a neural network to store data, all cat images would be stored in one localized area.
6. Learning all rules of life. This topic is very important. The most important game the robot must play is the game of life. All decision making pathways in the robot’s brain are structured to follow a very complex law system. Not only can’t the robot kill or harm a human being, he has to follow all laws and principals of the United States. This includes things like, the robot can’t shoplift, or rob a bank, or run naked out on the streets.
All etiquette of life must be learned. If the robot needs to attend a party, he needs to know: how to dress, how to socialize, when to eat, how to behave, and what to bring to the party. The only way to do that is to send the robot to school, from kindergarten to college, to learn the complicated game of life.
video on recursive planning:
video on natural language processing:
The content in this video was disclosed to the public, via patents and published books, between 2006-2007. So, this isn’t a recently discovered technology. It has been in the public domain for over 11 years.
The topics above are the major obstacles facing Artificial Intelligence in 2017. I solved these obstacles a long long long time ago. The proof is my patent applications filed.