用检索增强的多提示集成，高效提取小粒作物育种文献中的信息。
Retrieval-Augmented Multi-Prompt Ensemble for Minor-Grain Breeding Information Extraction
- 通过检索+多提示策略，不训练模型也能生成多样输出。
- 在12类实体和6类关系上总分0.499，比GPT-5.5提升11.4%。
- 适合需要高精度、无训练成本的信息抽取场景。
本文针对CCL2026-Eval Task 5：小粒作物育种信息抽取（MGBIE）任务，提出RAME（检索增强的多提示集成）框架。该系统无需训练，从育种文献中联合抽取12类实体与6类关系。RAME结合三阶段设计：(i) 基于混合BM25-嵌入检索器的少样本检索增强选择；(ii) 三种提示（严格、宽松、平衡）覆盖精确率到召回率的完整谱系；(iii) 大规模重复采样与多数投票机制过滤噪声预测。基于DeepSeek-V4-Flash模型，RAME在排行榜取得总分0.499（NER 0.730，RE 0.346），排名第一，超越官方基线GPT-5.5（0.448），相对提升11.4%。代码已开源。
原文摘要 · Abstract (English)
This paper presents our system for CCL2026-Eval Task 5: Minor-Grain Breeding Information Extraction (MGBIE), which jointly extracts 12 entity types and 6 relation types from minor-grain breeding literature. We propose RAME (Retrieval-Augmented Multi-Prompt Ensemble), a training-free framework that elicits multiple LLM outputs under controlled diversity and aggregates them by majority voting to obtain high-confidence predictions. RAME combines (i) retrieval-augmented few-shot selection via a hybrid BM25-embedding retriever, (ii) a three-prompt ensemble (Strict, Relaxed, Balanced) spanning the precision to recall spectrum, and (iii) large-scale repeated sampling with majority voting to filter noisy predictions. Built on DeepSeek-V4-Flash, RAME achieves a Total Score of 0.499 (NER 0.730, RE 0.346) on the leaderboard, ranking 1st and surpassing the official Track-A baseline powered by GPT-5.5 (0.448), representing an 11.4% relative improvement. Code is available at https://github.com/king-wang123/CCL26-RAME.
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。