让大模型先分清必须满足和可选需求，避免乱答错答。
First Things First: Teaching LLM-Based Agents to Prioritize Must-Haves before Nice-to-Haves
- 设计新强化学习方法，强制模型优先处理必须满足的需求。
- 在3649个真实场景任务中，成功率显著高于现有模型。
- 适用于电商、订票等复杂需求场景，适合做智能助手研发者参考。
多模态大语言模型（MLLM）在作为自主代理执行现实任务方面展现出巨大潜力，但针对用户复杂结构化需求的任务仍研究不足。本文考察三种需求场景：(i) 必须满足的需求唯一确定可行解；(ii) 多个答案满足必须需求，按可选需求排序；(iii) 无解满足必须需求，应拒绝回答。我们在3,649个精心构建的问题上评估了当前最先进的MLLM，涵盖电商、预订、地图与打车等真实服务场景，结果发现现有模型在所有场景均出现灾难性失败：误解需求、违反必须条件、生成无效解。为此，我们提出First Things First Reinforcement Learning（FTF-rl），显式优化对多优先级需求的推理能力。实验表明，该方法显著提升任务成功率，且在LogicVista、MathVision、InfoQA等常见逻辑与数学推理任务中也表现出泛化效果。研究显示，增强需求感知推理是提升MLLM代理泛化能力的简洁有效路径。代码与数据集见https://github.com/claire62/FTF-RL。
原文摘要 · Abstract (English)
Recent progress in multimodal large language models (MLLMs) has fueled significant enthusiasm in their potential to act as autonomous agents for real-world tasks. However, scenarios requiring agents to fulfill users' complex, structured requirements remain largely underexplored. In this work, we examine reasoning tasks under three distinct requirement scenarios: (i) Must-have requirements uniquely determine a unique feasible solution; (ii) Multiple answers satisfy the must-have requirements and are prioritized via the nice-to-have requirements; and (iii) No candidate solution satisfies the must-have requirements, in which case the agent should abstain from generating a response. We evaluate state-of-the-art MLLMs on 3,649 carefully constructed problems that reflect realistic service scenarios, including e-commerce, booking, and map-based or ride-hailing. Our evaluation reveals that existing MLLMs exhibit catastrophic failures in all scenarios. They frequently misinterpret task requirements, violate must-have requirements, and produce invalid solutions. To address this critical gap, we propose First Things First Reinforcement Learning FTF-rl that explicitly optimizes reasoning over multi-priority user requirements. Experimental results show that our method substantially improves the task success rate compared to strong baselines. Moreover, FTF-rl yields general effectiveness on popular logical and mathematical reasoning tasks, including LogicVista, MathVision, and InfoQA. Our findings suggest that enhancing requirement-aware reasoning capability provides a simple yet effective pathway to improve generalization of MLLM agents. Code and dataset are available at https://github.com/claire62/FTF-RL.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。