UTE is meant to be an engine for “theories that get better”. Experiment planning turns uncertainty into action: choose measurements
or interventions that reduce uncertainty (or disambiguate mechanisms) while respecting constraints.
What exists today
- Planning and CSP solving: DS16
- Core verb/operator patterns for actions: DS07g
UTE requirement: experiments are information actions
Objective
Define information objectives (reduce entropy, distinguish hypotheses) as first-class planning goals.
Constraints
Plan under real constraints: cost, time, feasibility, ethics, safety.
Closed loop
Execute the experiment, ingest results as evidence, revise the model, and re-plan.
Proposed extensions (research direction)
- Experiment operators with explicit preconditions, effects, and evidence output schemas.
- Information gain estimation integrated with uncertainty semantics.
- Batch planning for multi-step experimental protocols.
Spec
This page is summarized and formalized in DS38.