#AIQI#Model-Free AI#Reinforcement Learning#Universal Induction#Asymptotic Optimality#AIXI#Action-Value Functions#Artificial Intelligence Research
📌 Key Takeaways
AIQI is the first model-free agent proven asymptotically ε-optimal in general reinforcement learning
Previous optimal agents like AIXI were all model-based, requiring explicit environment models
AIQI performs universal induction over distributional action-value functions instead of policies or environments
Under a 'grain of truth condition,' AIQI achieves strong asymptotic ε-optimality and ε-Bayes-optimality
This breakthrough significantly expands the diversity of known universal agents
📖 Full Retelling
Researchers Yegon Kim and Juho Lee introduced AIQI (Universal AI with Q-Induction), the first model-free artificial intelligence agent proven to be asymptotically ε-optimal in general reinforcement learning, in a paper submitted to the arXiv preprint repository on February 26, 2026, addressing the limitation that all previous optimal AI agents were model-based. The breakthrough research presents a novel approach where AIQI performs universal induction over distributional action-value functions, diverging from previous works that focused on policies or environments. This significant advancement expands the diversity of known universal agents and establishes new theoretical foundations for artificial intelligence systems that can learn and adapt without explicit environmental models. Under a 'grain of truth condition,' the researchers have mathematically proven that AIQI achieves strong asymptotic ε-optimality and asymptotic ε-Bayes-optimality, placing it on par with the most advanced model-based agents like AIXI. The paper's abstract highlights that this development challenges the conventional wisdom in reinforcement learning, where optimal agents were believed to necessarily require explicit environment models.
🏷️ Themes
Artificial Intelligence, Reinforcement Learning, Machine Learning Theory
Mathematical formalism for artificial general intelligence
AIXI is a theoretical mathematical formalism for artificial general intelligence.
It combines Solomonoff induction with sequential decision theory.
AIXI was first proposed by Marcus Hutter in 2000 and several results regarding AIXI are proved in Hutter's 2005 book Universal Artificial Intelligence.
In machine learning and optimal control, reinforcement learning (RL) is concerned with how an intelligent agent should take actions in a dynamic environment in order to maximize a reward signal. Reinforcement learning is one of the three basic machine learning paradigms, alongside supervised learnin...
--> Computer Science > Artificial Intelligence arXiv:2602.23242 [Submitted on 26 Feb 2026] Title: A Model-Free Universal AI Authors: Yegon Kim , Juho Lee View a PDF of the paper titled A Model-Free Universal AI, by Yegon Kim and 1 other authors View PDF Abstract: In general reinforcement learning, all established optimal agents, including AIXI, are model-based, explicitly maintaining and using environment models. This paper introduces Universal AI with Q-Induction , the first model-free agent proven to be asymptotically $\varepsilon$-optimal in general RL. AIQI performs universal induction over distributional action-value functions, instead of policies or environments like previous works. Under a grain of truth condition, we prove that AIQI is strong asymptotically $\varepsilon$-optimal and asymptotically $\varepsilon$-Bayes-optimal. Our results significantly expand the diversity of known universal agents. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2602.23242 [cs.AI] (or arXiv:2602.23242v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2602.23242 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Yegon Kim [ view email ] [v1] Thu, 26 Feb 2026 17:21:16 UTC (141 KB) Full-text links: Access Paper: View a PDF of the paper titled A Model-Free Universal AI, by Yegon Kim and 1 other authors View PDF TeX Source view license Current browse context: cs.AI < prev | next > new | recent | 2026-02 Change to browse by: cs References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media C...
