Preprint / scholarly article
Natural Law-Type Conditions for Intelligent Self-Modifying Systems under Local Observation
- Published
- DOI
- 10.5281/zenodo.17860394
Abstract
The setting assumes only local observation: at each step the system has access solely to its current state, with no privileged meta-level, external clock, or separate optimisation oracle.
Keywords
- no-meta
- self-improvement
- self-modifying systems
- local observation
- evaluation functional
- lyapunov stability
- markov chains and stochastic stability
- intelligent region
- safe reinforcement learning
- AI safety
- AI alignment
- value-based reinforcement learning