Cognition · Software component

Test-Based Vote Aggregator

Software componentCognitionCognition & Memoryarc:TestBasedVoteAggregator

A self-consistency aggregator for generated code that executes each sampled implementation against provided test cases and selects among the samples that pass, instead of comparing output text.

Responsibility. Selects consensus code solutions by test-execution outcomes.

Also known as: Test-driven aggregation, Execution-based voting

Variant of Self-Consistency Aggregator abstract

When to choose. Choose for code generation where implementations vary textually but test cases define correctness.

Direct neighbourhood (hover for relationship types)

Relationships

invokes dependency

Code Execution Runner Ch5.3

reads dependency

Unit Test Suite Ch5.3

alternative to variability

Quantitative guidance

As stated by the sources; verify before use.

k=5 with consensus selection improved HumanEval correctness by 28%; 76% vs 59% of internal tasks passed all tests; consensus code rated more readable 64% of the time (Ch5.3).

Classification

Patterns: Test-based voting
Quality attributes: Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)Maintainability (ISO/IEC 25010)
Risks mitigated: Syntactically invalid generated codeEdge-case logic errors

Sources

Ch5.3: T. Nguyen, "Self-Consistency Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.3. ISBN: 9798244538229.