FusionMaestro: Harmonizing Early Fusion, Late Fusion, and LLM Reasoning for Multi-Granular Table-Text Retrieval

Sungho Park, Joohyung Yun, Jongwuk Lee, Wook-Shin Han

Keywords: Table-text retrieval, Information retrieval, Open-domain, Large language model, Early fusion, Late fusion, Multi-granular

Abstract: Table-text retrieval aims to retrieve relevant tables and text to support open-domain question answering. Existing studies use either early or late fusion, but face limitations. Early fusion pre-aligns a table row with its associated passages, forming ``stars," which often include irrelevant contexts and miss query-dependent relationships. Late fusion retrieves individual nodes, dynamically aligning them, but it risks missing relevant contexts. Both approaches also struggle with advanced reasoning tasks, such as column-wise aggregation and multi-hop reasoning. To address these issues, we propose FusionMaestro, which combines the strengths of both approaches. First, the edge-based bipartite subgraph retrieval identifies finer-grained edges between table segments and passages, effectively avoiding the inclusion of irrelevant contexts. Then, the query-relevant node expansion identifies the most promising nodes, dynamically retrieving relevant edges to grow the bipartite subgraph, minimizing the risk of missing important contexts. Lastly, the star-based LLM refinement performs logical inference at the star subgraph rather than the bipartite subgraph, supporting advanced reasoning tasks. Experimental results show that FusionMaestro outperforms state-of-the-art models with a significant improvement up to 42.6% and 39.9% in recall and nDCG, respectively, on the OTT-QA benchmark.

Primary Area: applications to computer vision, audio, language, and other modalities

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 10598