ChemWorld: Programmable Chemical Worlds for Controlled and Replayable Agent Experimentation

2026-08-11

AI Summary

This work addresses the challenge that existing chemical experimentation environments struggle to systematically manipulate underlying chemical rules while maintaining consistent task interfaces and lack reproducible, auditable interaction traces. To overcome this, the paper introduces ChemWorld—a programmable chemical environment that, for the first time, decouples public experimental interfaces from private chemical laws through modular components, a transactional execution engine, and a world-compilation mechanism. This design enables targeted intervention on individual hidden rules under fixed tasks and guarantees full replayability via transaction-level execution logs. The system validates the complete lifecycle of 52 formulations and 8 deterministic cases, effectively isolates single-variable effects across six parent–child world pairs, and demonstrates that agents can autonomously conduct end-to-end experiments in non-reference worlds.

Abstract

Autonomous chemistry increasingly depends on environments in which agents can repeatedly act, observe, and adapt.Physical laboratories provide essential real-material evidence but are costly to repeat and difficult to use for tightly matched interventions, whereas most digital environments keep the underlying experimental world largely fixed. We introduce ChemWorld, a programmable chemical environment in which reusable process and observation components are compiled into executable worlds. ChemWorld separates the public experimental contract available to an agent from evaluator-owned chemical and material laws. Researchers can therefore vary world composition and operating conditions, or change a single hidden law while holding the public task and interaction conditions fixed. Transactional execution records operations, failures, resource changes, and state transitions, allowing complete environment-action trajectories to be replayed exactly and audited. Full-census qualification covered the reference registry, 52 generated compositions, and module, interface, compilation, and invalid-action tests. Eight deterministic experimental cases demonstrated shared lifecycle semantics, failure recovery, and exact replay, while six parent-child world-fork pairs isolated the effects of single private-law interventions under matched public conditions. An independent agent also completed a full lifecycle in a non-reference world through the same public interface. Within the declared component and model domain, ChemWorld provides a controlled and replayable substrate for studying experimentation across systematically varied chemical worlds, complementary to physical-laboratory evidence and calibration.

Problem

• programmable chemical environments
• controlled experimentation
• replayable agent trajectories
• systematic intervention
• autonomous chemistry

Innovation

• programmable chemical environment
• replayable experimentation
• agent-environment interface
• transactional execution
• systematic world variation