Verify modular agents by scoring claims not task outcomes
Verify Claims, Not Scores: Evidence-Based Verification of Modular Agents
Artificial IntelligenceMachine Learning
Summary
When developers change parts of an AI agent, like its decision-making or checking system, they usually look at a single overall score to see if it got better. But that score doesn't explain why it changed or which part helped or hurt. The authors suggest checking each claim the agent makes, plus the evidence supporting it, giving a clearer picture of what's really working. They tested this on a financial AI agent and found some surprising results about what parts mattered and what didn't.
What this means in practice
- •For ai system engineers: Identify which specific components of modular AI agents improve or degrade performance by auditing claims rather than relying on aggregate scores.
- •For financial algorithm developers: Evaluate portfolio allocation agents more precisely by verifying component claims about hidden market regimes and decision boundaries.
Tested on simulated data.
Authors
Ali Atiah Alzahrani
Abstract
When developers change one component of an agent, such as its controller, a learned model or its verifier, they usually judge the change by an aggregate task score. That score cannot tell whether improvement was attainable, which component lost value, or what the agent's own checks certify. We introduce a claim-specific verification audit for modular agents that plan, act, check and refine. Instead of scoring the agent, the audit scores the evidence: each conclusion is recorded with the evidence behind it, one of four verdicts (supported, unsupported, unresolved or not evaluated) and the boundary within which it holds. Three tools supply that evidence. Oracle policies measure attainable improvement under an explicitly stated action set, so that a low value can be traced to the evaluation rather than to the environment. Replacing one component at a time with a perfect counterpart locates lost value, with null results read as unresolved whenever a downstream component could mask them. A separate test asks whether the verifier's score identifies the quantity it is read as bounding. Applied to a constrained portfolio-allocation agent in a synthetic market with known hidden regimes, the audit shows that the value of perfect regime information depends on the action set used to measure it, that the scenario generator discards most of the regime signal while better local fidelity does not improve decisions, and that the runtime verifier can be bypassed with no visible change in outcomes. The contribution is the protocol and the evidential distinctions it enforces; the empirical findings are specific to the agent and environment studied.