用合成对比推理轨迹提升多表问答模型的推理能力
Synthetic Contrastive Reasoning for Multi-Table Q&A
- 构建正负例推理轨迹对，通过对比优化训练模型
- 在多个大模型上实现最高21个百分点的性能提升
- 适合需要强逻辑推理的多表问答研究与应用
多表问答需模型检索相关证据、关联模式并跨关系表进行组合推理。现有资源通常只提供问题和答案，缺乏推理过程监督。为此，我们利用异构大模型生成经验证的正向推理轨迹和合理负向轨迹，构建合成对比推理轨迹数据集（MMQA）。基于该数据集，采用对比偏好优化（CPO）微调开源大模型。在Qwen3-14B、Mistral-8B和Llama-3.1-8B上，相比传统问答监督微调，平均提升9.7%至16.3%，在MMQA上最高提升达21个百分点。消融实验表明，异构生成器增强了对比信号；自动与人工评估均显示生成样本具备高度忠实性、连贯性和有意义的对比性。
原文摘要 · Abstract (English)
Multi-table question answering requires models to retrieve relevant evidence, link schemas, and perform compositional reasoning across relational tables. Existing multi-table Q&A resources typically provide questions and final answers but lack reasoning supervision that explains how answers are derived. To address this gap, we construct a synthetic contrastive reasoning-trace dataset for MMQA by generating validated positive traces and plausible negative traces with heterogeneous LLMs. We then use the resulting preference pairs to fine-tune open-weight LLMs with Contrastive Preference Optimization (CPO). Across Qwen3-14B, Mistral-8B, and Llama-3.1-8B, CPO achieves absolute average improvements over Q&A supervised fine-tuning ranging from 9.7%-16.3%, with gains up to 21 percentage points on MMQA. Ablations show that heterogeneous positive and negative trace generators strengthen the contrastive signal, and automated as well as human evaluations indicate that the generated pairs are largely faithful, coherent, and meaningfully contrastive.
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。