OpenCogPrime:Roadmap

From OpenCog
Jump to: navigation, search

This is an old version of the OpenCog roadmap. For a more recent roadmap see the High-Level roadmap and the technical roadmap.

OpenCogPrime Roadmap

The overall long-term goal of the project described in this roadmap is the creation a self-modifying, self-improving software program with superhuman general intelligence and a beneficent ethical attitude. (No one ever accused us of not being ambitious enough!)

A high-level roadmap for OpenCogPrime development, in six phases:

The boundaries between the phases are not entirely crisp, but the breakdown into phases provides useful guidance. Phases 3-6 are yet to be described in detail.

Phase 1: Basic Components

Completion of essential aspects of AGI system design (mathematical, conceptual and software-architecture); and implementation of initial versions of key components.

Phase 2: Artificial Toddler

On this topic, see this paper on "AGI Preschool" that was submitted from AGI-09:

http://www.opencog.org/wiki/Image:Preschool.pdf

Refinement of design and implementation in the course of teaching the AI system to control an agent in a simulation world, according to a loosely Piagetan learning plan.

    • Goal: an “artificial toddler” with
      • qualitatively intelligent though not humanlike English conversation ability, involving simple sentences appropriately contextually deployed
      • the approximate problem-solving ability of an average four-year old human child within the context of its simulation world

Phase 3: Artificial Child

Interaction with the "artificial toddler" so as to teach it to more effectively think and communicate.

    • Goal: an “artificial child” with the approximate problem-solving and communicational ability of an average ten-year old human child within the context of its simulation world

Phase 4: Artificial Adult

Instruction of “artificial child” in relevant topics, with a focus on bioscience, mathematics and ethics. Refinement of implementation as necessary.

    • Goal: an intelligent, ethical “artificial adult” and young “artificial scientist”

Phase 5: Artificial Scientist

Instruction of artificial scientist in AI design and general computer science.

    • Goal: an ethical AI capable of radically modifying and improving its own implementation in accordance with its goals

Phase 6: Artificial Intellect

An AI created by an artificial scientist.

    • Goal: an ethical intellect capable of managing the AI scientists

Status

In terms of the above breakdown, at present we are near the start of Phase Two, and still wrapping up some aspects of Phase One.

Most of Phase One was completed prior to the inception of the OpenCogPrime project as such. And it is, of course, the near-completion of Phase One that allows us to make a serious roadmap and task breakdown for Phase Two.

The conceptual and mathematical aspects of Phase One are embodied in the OpenCogPrime wikibook and related documents.

The implementation aspect of Phase One is underway, and largely involves the porting-over to OpenCog of (simplified, cleaned-up versions of) components from the Novamente Cognition Engine. But it also involves some original R&D, such as the development of the initial attention allocation code that Joel Pitt has worked on.

Project Phases and Piagetan Stages

Following Piaget, a rough overview of the learning milestones involved in the above high-level plan is as follows:

  • Infant stage: object permanence, word-object associations, the basic ability to perceive, move around and grab
  • Pre-operational stage: simple syntax, the ability to build a variety of simple structures, simple one-dimensional classification of objects
  • Concrete-operational stage: understanding of conservation laws, complex syntax, the building of complex structures according to multi-part plans, modeling of others’ minds, complex classification of objects
  • Formal stage: abstract logical inference, scientific hypothesis and testing
  • Reflexive (post-formal) stage: attempt to objectively understand, modify and improve one's own self

Phase Two of the above high-level project plan requires the AI system to master the infant and pre-operational stages. Phase Three requires the concrete-operational stage, verging on parts of the formal stage. Obviously, Phase Four requires full-on, highly-sophisticated formal reasoning; and Phase Five corresponds to an intensive reflexive stage system.

For a more careful consideration of these Piagetan phases in the context of OpenCogPrime like AGI systems, see

http://www.novamente.net/file/WCCI06_Stages.pdf

and note also that these stages may be related to ethical as well as cognitive development, see

http://www.agiri.org/AGIethical.pdf