Overview

Bayesian Network Coverage-directed Test Generation is identified in the literature as a coverage-guided test generation approach for functional verification that uses Bayesian networks. The DATE 2022 paper Cross-Level Processor Verification via Coverage-guided Aging cites the original reference as S. Fine and A. Ziv, "Coverage directed test generation for functional verification using bayesian networks", published at DAC 2003.

Position in test-generation literature

In the DATE 2022 related-work discussion, Bayesian-network-based coverage-directed generation is presented as an alternative to model-based test generators that use constraint-based specifications. The paper distinguishes:

• model-based test generators that leverage constraint-based specification formats, and
• alternative approaches that include coverage-guided test generation based on Bayesian networks, other machine-learning techniques, fuzzing, and symbolic execution.

Limitations noted in the cited survey

The DATE 2022 survey states that this broader class of alternative approaches—including Bayesian-network-based coverage guidance—is either not designed for RTL verification or imposes restrictions on the generated instruction streams. It also states that these approaches do not target the modern RISC-V ISA.

Scope of this article

Based on the available evidence, the concept can be characterized at a high level as a coverage-guided functional-verification technique using Bayesian networks and as a non-model-based alternative in processor test-generation literature. The provided evidence does not describe the internal Bayesian-network formulation or algorithmic details further, so those details are omitted here.