Bayesian Exploration for LLM Agents
Why raising temperature is not curiosity, and how posterior sampling turns uncertainty into coherent, hypothesis-driven exploration
PlugBO: A Modular Framework for Agentic Bayesian Optimization
A modular, plug-and-play framework for agentic Bayesian optimization
World Models for Scientific Discovery
Why prediction alone is not discovery, and what world models need to support explanation, experimentation, and abduction
Metacognitive Self-Modification in Self-Improving Agents
How Hyperagents extends the Darwin-Gödel Machine by letting agents modify their own improvement process
Learning to Simulate and Act in the Physical World
How interactive world models are built, and how agents learn inside them
Exploration as a Path to General Intelligence
Why exploration, not just exploitation, may be the missing ingredient in current AI systems
The Quest for Open-Endedness in AI
Tracing open-endedness from cybernetics to modern open-ended learning systems
Do Agents Need a World Model?
From Ilya Sutskever's conjecture to DeepMind's formal argument for world models in general agents
Self-Improving Coding Agents
How the Darwin-Gödel Machine rewrites its own code to improve, and whether we can trust it
Algorithm Discovery with Large Language Models
How large language models search program space to discover new algorithms
A Unified View of Bayesian Optimization and Active Learning
Unifying Bayesian optimization and active learning as goal-driven adaptive sampling
No posts match that search.