我们能操控黑箱吗？迈向以可控性为中心的协同代理推荐系统评估

中文摘要 AI 辅助

推荐系统如黑箱般运作，用户和监管者无法引导其输出以符合特定意图或审查其行为。这种缺乏可控性（即系统响应明确指导的能力）在现有评估范式中仍是未解决的维度。为填补这一空白，我们提出CtrlBench-Rec，一个用于系统评估可控性的协同多智能体框架。我们将三个基本任务形式化：目标内容发现、兴趣轮廓塑造和流行度偏差缓解，它们共同衡量从显式命令到隐式表示引导以及最终克服算法偏差的可操纵性。在真实数据集和多个推荐模型上的实验表明，我们的框架有效量化了可控性并揭示了关键系统瓶颈，尤其是对引导长尾内容的持续阻力。CtrlBench-Rec为可控推荐研究、算法审计和用户赋权提供了首个标准化工具包。我们的代码已在该https URL上发布。

英文摘要

Recommender systems operate as Black-Boxes, leaving users and regulators unable to steer their outputs toward specific intentions or audit their behavior. This lack of controllability, defined as the system's ability to respond to explicit guidance, remains an unaddressed dimension in existing evaluation paradigms. To fill this gap, we propose CtrlBench-Rec, a collaborative multi-agent framework for systematic assessment of controllability. We formalize three fundamental tasks: target content discovery, interest profile shaping, and popularity bias mitigation, which together measure steerability from explicit commands to implicit representation steering and finally to overcoming algorithmic biases.Extensive experiments on real-world datasets and multiple recommendation models demonstrate that our framework effectively quantifies controllability and exposes critical system bottlenecks, most notably persistent resistance to guiding long tail content. CtrlBench-Rec provides the first standardized toolkit for controllable recommendation research, algorithmic auditing, and user empowerment. Our code is released on https://github.com/caskcsg/CtrlBenchRec.

发表机构

• Institute of Information Engineering, Chinese Academy of Sciences, China（中国科学院信息工程研究所）
• School of Cyber Security, University of Chinese Academy of Sciences, China（中国科学院大学网络安全学院）

机构由 AI 辅助整理，请以论文原文为准。