Agentic AI Reference Architecture

Cognition · Software component

Action Prior Estimator

Software componentCognitionCognition & Memoryarc:ActionPriorEstimator

A cognition component that serves policy-network action probabilities for newly expanded search nodes and learned rollouts, biasing exploration toward promising actions.

Responsibility. Supplies informed action priors to tree search and rollouts.

Also known as: Policy prior provider

is invoked byis invoked byhostsMCTS Planner: is invoked byMCTS PlannerRollout Simulator: is invoked byRollout SimulatorPolicy Network: hostsPolicy Network
Direct neighbourhood (hover for relationship types)

Relationships

hosts structural

is invoked by dependency

Classification

Patterns
Policy-network priorsLearned rollout policy
Risks mitigated
Wasting limited budget on implausible actions

Sources

  1. Ch5.5: T. Nguyen, "Monte Carlo Tree Search Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.5. ISBN: 9798244538229.