用轻量级结构桥接表格与大模型，提升多表问答性能。
Latent Bridges for Multi-Table Question Answering
- 构建异构图编码表结构，通过消息传递提取关系特征。
- 仅训练91M参数的图编码器和隐变量桥，大模型完全冻结。
- 在复杂多表场景下显著提效，适合需要结构化推理的场景。
我们提出GRAB，一种用于表格问答的构造-编码-桥接流水线。该方法将关系型数据转换为异构图，通过消息传递进行编码，并利用少量查询条件相关的隐变量令牌，将信息传递给大语言模型（LLM）。这一过程为LLM提供紧凑且任务相关的结构表示，同时保留原始文本。关键在于，大模型始终保持冻结状态，以保护其通用推理能力；仅训练轻量级图编码器和隐变量桥（共9100万参数），实现高效训练。该流水线在关系型问答任务中显著提升性能，尤其在复杂的多表场景下表现突出，为关系型深度学习与大模型的结合提供了高效且原理清晰的解决方案。
原文摘要 · Abstract (English)
We introduce GRAB, a constructor-encoder-bridge pipeline for table question answering. Our method lifts relational data into an heterogeneous graph, encodes it via message passing, and transfers the signals to an LLM through a small set of query-conditioned latent tokens. This provides the LLM with a compact, task-relevant structural representation together with the flattened text. Crucially, the LLM remains strictly frozen to preserve its general reasoning capabilities; we train only the lightweight graph encoder and latent bridge (91M parameters), allowing the entire pipeline to be trained efficiently. Our pipeline significantly improves performance on relational Question Answering, with the largest gains in demanding multi-table settings, offering an efficient, principled way to connect relational deep learning with LLMs.
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。