Discovery - Rei

Discovery lets Adapt-1 form task-specific structure from admitted evidence while the Domain defines the public task contract. An authored learner-facing ontology is optional.

Task contract and learned structure

The application defines the information available to Adapt-1 and the result it needs. The environment supplies observations and execution outcomes.

Application contract	Structure that can form from admitted evidence
Event and outcome meanings	Useful learner input paths
Target or result type	Predictive combinations of observable fields
Episode and reset boundaries	Bounded temporal relationships
Available actions and reward semantics when applicable	State-dependent action values and delayed-credit structure
Observable before/after event boundary	Transition and causal bindings where supported
Safety and execution constraints	Predictive rules and other task-local structure

Discovery paths

Discovery path	What forms from evidence
Transition Discovery	Executable input projections and compatible before/after causal bindings
Structure Discovery	Outcome-useful fields, combinations, bounded temporal lags, and predictive rules
Sequential Discovery	State-dependent action values and delayed credit across ordered interaction

A Domain can combine compatible paths under its declared learning lifecycle. Each path has a contract for targets and admissible evidence, including attribution and abstention behavior.

Learning schedules

Define the complete learning-run boundary when configuring the state lifecycle.

Schedule	Discovery behavior
Zero-start learning across a run	The complete run begins with empty task-specific learned state. Discovery forms structure from events inside the run and can retain it across compatible episodes or passes.
Zero-start task-local learning	Each independently redrawn task begins empty. Structure forms from local evidence and is reset after the task produces its result.
Separate acquisition, then frozen use	Discovery forms structure from acquisition records outside the new run. The acquired state is then used without learner writes.
Separate acquisition, then continued adaptation	Discovery forms the starting state outside the new run. Current outcomes continue revising it.

State retained between episodes belongs to a zero-start run when the complete run began empty.

Configure the ontology

Mode	When to use it
Discovery	Useful task-facing structure should form from experience under a defined public task contract
Authored ontology	The application requires task-facing structure to be explicitly reviewed or fixed before use
Mixed	Reviewed parts of the task-facing structure remain fixed while other supported structure is discovered

After acquisition, inspect the discovered structure and evaluate it on held-out work. You can pin reviewed parts of the ontology when the application requires a stable production contract.

Inspect Discovery state

Public API responses can expose Discovery readiness and active input paths. Depending on the configured learning method, diagnostics also include learned structure, supporting or contradicting evidence, learner versions, and abstention states. Use the returned diagnostics and held-out outcomes to evaluate the acquired state before relying on it.

Supply observations and feedback

Learning requires the observations and outcomes specified by the selected contract. Available actions are limited to those the application permits. Send only information available at the decision point; keep hidden answers, private evaluator state, and later outcome values out of pre-outcome inputs.

Choose a learning setup

Configure when state starts empty, carries forward, or freezes.

Use Discovery

Configure Discovery and inspect the structure formed from events.

Sequential Discovery

Form state-dependent action values and delayed credit from ordered interaction.

Author a Domain ontology

Supply reviewed task-facing structure when the application requires it.

Choose how a Domain learns

Choose how observations and outcomes should improve later results.

Adapt-1 foundations

Read how Adapt-1 uses retained state to produce structured results.

​Task contract and learned structure

​Discovery paths

​Learning schedules

​Configure the ontology

​Inspect Discovery state

​Supply observations and feedback