Interface: SynthesizeRequest

@kortexya/reasoninglayer

---

@kortexya/reasoninglayer / ILP / SynthesizeRequest

Interface: SynthesizeRequest

Defined in: src/types/ilp.ts:200

Request for neural-guided rule synthesis.

Remarks

Combines ILP with optional GFlowNet-guided sampling to synthesize rules from training examples.

Properties

examples

examples: TrainingExample[]

Defined in: src/types/ilp.ts:202

Training examples as input/output pairs.

---

maxProposals?

optional maxProposals: number | null

Defined in: src/types/ilp.ts:204

Maximum proposals to try.

---

targetSort

targetSort: string

Defined in: src/types/ilp.ts:206

Target sort name to synthesize rules for.

---

useGflownet?

optional useGflownet: boolean

Defined in: src/types/ilp.ts:208

Whether to use GFlowNet-guided sampling.