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

TL;DR

HELIOS introduces a novel multi-granular retrieval framework combining edge-based subgraph retrieval, node expansion, and LLM refinement to improve table-text retrieval accuracy and reasoning capabilities in open-domain QA.

Contribution

The paper proposes HELIOS, a hybrid approach that integrates early and late fusion techniques with LLM reasoning, addressing limitations of existing methods in multi-hop and complex reasoning tasks.

Findings

Achieves up to 42.6% recall improvement on OTT-QA.

Outperforms state-of-the-art models in nDCG by 39.9%.

Effectively supports advanced reasoning like multi-hop and column-wise aggregation.

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 HELIOS, 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 graph level rather than the bipartite subgraph, supporting advanced reasoning tasks. Experimental results show that HELIOS 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.