Reasoning-learning systems based on non-axiomatic reasoning system theory
Draft — This content has not been approved for publication.
Authors: Patrick Hammer
← Back to Publications
Reasoning-Learning Systems Based on Non-Axiomatic Reasoning System Theory
Author:
Year: 2022
Venue: Proceedings of Machine Learning Research, International Workshop on Self-Supervised Learning
Links: PDF
Summary
Compares a reasoning-learning approach based on Non-Axiomatic Reasoning System theory with common reinforcement-learning techniques. The paper frames ONA/NARS as a system for learning from experience under uncertainty while taking explicit background knowledge into account.
Relevance to Hyperon
This paper supports MeTTa-NARS Deep Dive by documenting the reasoning-learning side of the NARS/ONA tradition, especially the relationship between uncertainty reasoning, practical reasoning, and procedure learning.