ARGUS: Policy-Adaptive Ad Governance via Evolving Reinforcement with Adversarial Umpiring

TL;DR

ARGUS is a system for online ad governance that adaptively manages evolving policies using multi-agent adversarial reasoning and reinforcement learning, effectively handling label inconsistencies and regulatory changes.

Contribution

It introduces a three-stage framework with adversarial label rectification and knowledge discovery to improve policy adaptation in dynamic regulatory environments.

Findings

ARGUS outperforms traditional fine-tuning methods on industrial and public datasets.

The system effectively resolves conflicts between stale labels and new policies.

ARGUS achieves superior policy adaptation with minimal gold data.

Abstract

Online advertising governance faces significant challenges due to the non-stationary nature of regulatory policies, where emerging mandates (e.g., restrictions on education or aesthetic anxiety) create severe label inconsistencies and reasoning ambiguities in historical datasets. In this paper, we propose ARGUS, a policy-adaptive governance system that enables evolving reinforcement through multi-agent adversarial umpiring. ARGUS addresses the sparsity of new policy data by employing a three-stage framework: (1) Policy Seeding for initial perception; (2) Adversarial Label Rectification, which utilizes a ``Prosecutor-Defender-Umpire'' architecture to resolve conflicts between stale labels and new mandates; and (3) Latent Knowledge Discovery, which employs a tripartite dialectical discussion to unearth sophisticated, ``gray-area'' violations. By leveraging RAG-enhanced policy knowledge and Chain-of-Thought synthesis as dynamic rewards for reinforcement learning, ARGUS synchronizes its reasoning pathways with evolving regulations. Extensive experiments on both industrial and public datasets demonstrate that ARGUS significantly outperforms traditional fine-tuning baselines, achieving superior policy-adaptive learning with minimal gold data.