Framework unifies iterative policy improvement and recursive self-improvement
Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement
Artificial IntelligenceMachine Learning
Summary
Improving how intelligent agents learn over time can happen in different ways, but there hasn’t been one clear model to explain all types. The authors created a framework called Generalized Agent Iteration (GAI) that describes both traditional learning methods and the idea of agents improving themselves. They define agents by parts that can change and show how learning cycles involve checking and upgrading these parts. This new model helps compare current systems and better understand how agents can improve either with help from outside or by self-guidance.
What this means in practice
- •For autonomous system developers: Design iterative learning agents that improve their own mechanisms within the system boundaries.
- •For software architects: Compare and classify existing AI systems based on their improvement processes using a shared framework.
A position paper. It proposes an approach and reports no results.
Authors
Hongyao Tang, Yi Ma, Pengyi Li, Yifu Yuan
Abstract
When we speak of recursive self-improvement (RSI), are we speaking of a phenomenon, a mechanism, or a prospect? Towards autonomous and evolving intelligence, RSI is being claimed at many scales, while no single framework that formally describes these emerging instances exists. Its counterpart in the classical realm, iterative policy improvement, is characterized by generalized policy iteration (GPI), a framework of broad applicability with well-understood theoretical properties, but only where the update principle and the evaluation base lie outside the agent. In this paper, we propose Generalized Agent Iteration (GAI), a formal framework that describes iterative policy improvement and RSI as two cases of a single learning paradigm. GAI defines the agent as a configuration of modifiable components within a system and models the learning process as a cycle of agent evaluation and agent improvement. Two pivotal dials then distinguish the instances: whether the improving mechanism is part of the agent and whether the standard it is measured against is grounded outside it. The former dial delineates the boundary between GPI and RSI, and the latter determines a system's polarity as anchored, goal drift, or fully self-referential. Moreover, we use these coordinates to place existing systems on the same two axes and make the defects of recursive self-improvement statable one condition at a time. We see this paper as a first step toward exploring a formal characterization of RSI that rests on the classical account, makes existing systems comparable, and provides a principled basis for analyzing and designing new ones.