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 multihop 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.

Multimodal document retrieval aims to retrieve query-relevant components from documents composed of textual, tabular, and visual elements.An effective multimodal retriever needs to handle two main challenges: (1) mitigate the effect of irrelevant contents caused by fixed, single-granular retrieval units, and (2) support multihop reasoning by effectively capturing semantic relationships among components within and across documents.To address these challenges, we propose LILaC, a multimodal retrieval framework featuring two core innovations.First, we introduce a layered component graph, explicitly representing multimodal information at two layers-each representing coarse and fine granularity-facilitating efficient yet precise reasoning.Second, we develop a lateinteraction-based subgraph retrieval method, an edge-based approach that initially identifies coarse-grained nodes for efficient candidate generation, then performs fine-grained reasoning via late interaction.Extensive experiments demonstrate that LILaC achieves state-ofthe-art retrieval performance on all five benchmarks, notably without additional fine-tuning.We make the artifacts publicly available at github.com/joohyung00/lilac.

The schemalessness, one of the major advantages of JSON representation format, comes with high penalties in querying and operations by denying various critical functions such as query optimizations, indexing, or data verification. There have been continuous efforts to develop an accurate JSON schema discovery algorithm from a bag of JSON documents. Unfortunately, existing schema discovery techniques, being top-down algorithms, face challenges from the lack of visibility into children nodes of JSON tree. With absence of the information about lower-level JSON elements, top-down algorithms need to employ assumptions and heuristics to decide the schema type of nodes. However, such static decisions are often violated in datasets which causes top-down algorithms to perform poorly. To overcome this, we propose an algorithm, called ReCG, that processes JSON documents in a bottom-up manner. It builds up schemas from leaf elements upward in the JSON document tree and, thus, can make more informed decisions of the schema node types. In addition, we adopt MDL (Minimum Description Length) principles systematically while building up the schemas to choose among candidate schemas the most concise yet accurate one with well-balanced generality. Evaluations show that our technique improves the recall and precision of found schemas by as high as 47%, resulting in 46% better F1 score while also performing 2.11× faster on average against the state-of-the-art.