Cognition · Data artifact

Thought Search Policy

Data artifactCognitionCognition & Memoryarc:ThoughtSearchPolicy

A configuration artifact fixing tree-search parameters: breadth b, branching factor k, maximum depth and the pruning value threshold below which branches are abandoned.

Responsibility. Bounds search breadth, depth and pruning aggressiveness.

Also known as: Pruning policy, Pruning threshold, Search budget parameters

Direct neighbourhood (hover for relationship types)

Relationships

is written by dependency

Confidence Calibration Analyzer Ch5.2

constrains control

Tree Search Controller abstract Ch5.2

Design guidance

SHOULD limit branching factor k to 3-5 rather than ten or more alternatives per state.
SHOULD configure backtracking conservatively, using pruning thresholds to eliminate obviously unproductive branches early.
SHOULD adjust thresholds from measured value-estimate calibration.

Quantitative guidance

As stated by the sources; verify before use.

Example pruning threshold: rating 'impossible' or value < 3/10 (Ch5.2).

Classification

Patterns: Threshold pruningExploration-exploitation trade-off
Risks mitigated: Excessive backtracking costPruning of viable paths

Sources

Ch5.2: T. Nguyen, "Tree-of-Thought (ToT) Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.2. ISBN: 9798244538229.