通过约束检索与选择性辩论，提升段落级论点挖掘准确率
LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining
- 先用元数据检索缩小标签候选范围，再用维度上限约束解码
- 仅对不确定样本启动三智能体辩论，提升输出可靠性
- 适合关注结构化生成与小模型高效推理的研究者
我们提出 LLM-INSTRUCT，该系统在 UZH 共享任务 2026 年的联合国与教科文组织决议段落级论点挖掘任务中获得第一名。任务要求进行段落类型分类、预测 141 个官方标签子集，并在严格 JSON 模式下预测有向关系，仅使用最大 8B 参数的开源模型。系统将任务建模为受限结构化预测：首先利用元数据感知稠密检索缩小候选标签空间，再通过每维上限的约束解码，仅将不确定性案例转入三智能体辩论分支，并最终验证输出模式。在官方排行榜上，系统综合排名第一，F1 排名第一，LLM-as-a-Judge 排名第五。开发过程中，配置搜索使任务 1b 微平均 F1 从 35.83% 提升至 40.08%，同时保持任务 2 内部得分 4.421。核心经验是：生成前缩减决策空间可显著提高准确率与提交鲁棒性。代码与辅助脚本已公开于 https://github.com/LLM-Instruct-at-UZH-Shared-Task-2026/Method。
原文摘要 · Abstract (English)
We present LLM-INSTRUCT, the winning system for the UZH Shared Task at ArgMining 2026 on paragraph-level argument mining in UN and UNESCO resolutions. The task requires paragraph-type classification, prediction of a subset of 141 official tags, and directed relation prediction under a strict JSON schema setting using only open-weight models up to 8B parameters. We frame the task as constrained structured prediction. The system first narrows the candidate tag space with metadata-aware dense retrieval, then applies constrained decoding with per-dimension caps, escalates only uncertain cases to a three-agent debate branch, and finally validates the output schema. On the official leaderboard, LLM-INSTRUCT ranked 1st overall, with 1st in F1 and 5th in LLM-as-a-Judge. During development, our configuration search further improved Task 1b Micro-F1 from 35.83% to 40.08% while keeping the internal Task 2 score at 4.421. The main lesson is simple: reducing the decision space before generation improves both accuracy and submission robustness. Our code and supporting scripts are publicly available at: https://github.com/LLM-Instruct-at-UZH-Shared-Task-2026/Method
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。