Generic AI

From AGIRI.org

Contents 1 Introduction 2 Business model 3 Design features 4 Architecture 5 Participation 6 See also 7 External links

Introduction

Genifer is a full-scale AGI project.

Business model

Genifer is opensource but we plan to explore commercial opportunities. Contributors to Genifer will be given company shares.

Design features

Genifer:

1. is primarily logic-based (first-order predicate logic)
2. has a unified architecture (uses logic as the universal KR)
3. employs neural networks in low-level visual processing
4. augmented with probabilistic/fuzzy logic
5. has a modular architecture with relatively few (< 10) modules

Architecture

The Genifer architecture contains 8 basic modules:

1. deduction
2. pattern recognition
3. abduction
4. natural language
5. belief revision
6. planning / reinforcement learning
7. inductive learning
8. vision

Participation

Currently Genifer is under seed-stage development. If you want to get involved please contact YKY at General.Intelligence@Gmail.com.

See also

• Yan King Yin

External links

• Genifer source code hosted on Launchpad

• Our collaborative e-book on AGI theory

• General Intelligence Research Group (older web site, will soon be replaced)

• Genifer mailing list