PolyFrame在MWE-2026 AdMIRe 2：当词语不够时：多模态成语消歧
PolyFrame at MWE-2026 AdMIRe 2: When Words Are Not Enough: Multimodal Idiom Disambiguation
AI总结：
本文提出PolyFrame系统，通过统一管道处理图像+文本排名和纯文本描述排名任务，利用轻量模块提升多语言成语消歧性能，无需微调大模型。
摘要：
多模态模型在处理成语时面临非组合性含义的挑战，特别是在多语言环境下更为严峻。我们介绍了PolyFrame，用于MWE-2026 AdMIRe2共享任务的多模态成语消歧系统，包含统一的图像+文本排名（子任务A）和纯文本描述排名（子任务B）流程。所有模型变体保留冻结的CLIP风格视觉-语言编码器和多语言BGE M3编码器，仅训练轻量模块：逻辑回归和LLM基于的句子类型预测器、成语同义替换、干扰项感知评分以及Borda排名融合。从CLIP基线（英语开发集26.7% Top-1，英语测试集6.7%）开始，添加成语意识改写和显式句子类型分类后，性能提升至英语Top-1 60.0%，并在零样本迁移至葡萄牙语时达到60.0% Top-1（0.822 NDCG@5）。在多语言盲测中，我们的系统在15种语言上实现了子任务A平均Top-1/NDCG分数为0.35/0.73，子任务B为0.32/0.71。消融结果表明，成语意识改写是性能的主要贡献者，而句子类型预测和多模态融合增强了鲁棒性。这些发现表明，无需微调大型多模态编码器即可实现有效的成语消歧。
英文摘要：
Multimodal models struggle with idiomatic expressions due to their non-compositional meanings, a challenge amplified in multilingual settings. We introduced PolyFrame, our system for the MWE-2026 AdMIRe2 shared task on multimodal idiom disambiguation, featuring a unified pipeline for both image+text ranking (Subtask A) and text-only caption ranking (Subtask B). All model variants retain frozen CLIP-style vision--language encoders and the multilingual BGE M3 encoder, training only lightweight modules: a logistic regression and LLM-based sentence-type predictor, idiom synonym substitution, distractor-aware scoring, and Borda rank fusion. Starting from a CLIP baseline (26.7% Top-1 on English dev, 6.7% on English test), adding idiom-aware paraphrasing and explicit sentence-type classification increased performance to 60.0% Top-1 on English and 60.0% Top-1 (0.822 NDCG@5) in zero-shot transfer to Portuguese. On the multilingual blind test, our systems achieved average Top-1/NDCG scores of 0.35/0.73 for Subtask A and 0.32/0.71 for Subtask B across 15 languages. Ablation results highlight idiom-aware rewriting as the main contributor to performance, while sentence-type prediction and multimodal fusion enhance robustness. These findings suggest that effective idiom disambiguation is feasible without fine-tuning large multimodal encoders.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
