Philip C.Jackson

Toward Human-Level Artificial Intelligence

Notify me when the book’s added
To read this book, upload an EPUB or FB2 file to Bookmate. How do I upload a book?
How can human-level artificial intelligence be achieved? What are the potential consequences? This book describes a research approach toward achieving human-level AI, combining a doctoral thesis and research papers by the author.
The research approach, called TalaMind, involves developing an AI system that uses a 'natural language of thought' based on the unconstrained syntax of a language such as English; designing the system as a collection of concepts that can create and modify concepts to behave intelligently in an environment; and using methods from cognitive linguistics for multiple levels of mental representation. Proposing a design-inspection alternative to the Turing Test, these pages discuss 'higher-level mentalities' of human intelligence, which include natural language understanding, higher-level forms of learning and reasoning, imagination, and consciousness. Dr. Jackson gives a comprehensive review of other research, addresses theoretical objections to the proposed approach and to achieving human-level AI in principle, and describes a prototype system that illustrates the potential of the approach.
This book discusses economic risks and benefits of AI, considers how to ensure that human-level AI and superintelligence will be beneficial for humanity, and gives reasons why human-level AI may be necessary for humanity's survival and prosperity.
This book is currently unavailable
482 printed pages
Original publication



    How did you like the book?

    Sign in or Register

On the bookshelves

Drag & drop your files (not more than 5 at once)