Experiments on three real-world datasets demonstrate that AnnoIndex consistently outperforms state-of-the-art baselines, achieving the highest average F1 score while maintaining robust performance on complex multi-hop join and progressive reasoning queries.
Abstract
Unstructured documents constitute the majority of enterprise and web data. With the rapid development of large language models(LLMs), researchers have started to build data systems that analyze unstructured textual documents like operating on databases. However, because mainstream retrieval methods still relies on fuzzy matching based on vector similarity, accurately obtaining information and performing structured analysis and reasoning remains a major challenge. To address these limitations, AnnoIndex introduces two core fundamental components. The first is Annotation Index. The system uses a module called SchemaLoop to automatically create hierarchical annotation schemas from the raw corpus, and then uses lightweight language model to extract specific values. It turns scattered unstructured text into a materialized, structured index that enables low-cost filtering and querying. The annotation index avoids the black-box matching of vector similarity and amortizes attribute extraction costs from online queries to a one-time build. The second innovation is a Structured Query Engine. It compiles user questions into execution plans based on SQL extension. It first uses the Annotation Index for precise documents filtering, then gradually applies extraction operations in ascending order of cost, resorting to LLMs only for the remaining minimal fraction of the corpus that require deep semantic understanding. The extracted attributions are merged into the annotation index, reducing the cost of future queries. Experiments on three real-world datasets demonstrate that AnnoIndex consistently outperforms state-of-the-art baselines, achieving the highest average F1 score (0.87) while maintaining robust performance on complex multi-hop join and progressive reasoning queries.
The findings suggest that AI-based structured extraction may redefine how organisations formalise expertise, shifting from document-centric storage toward schema-driven knowledge architectures.
Dilyan Georgiev, E. Gourova· European Conference on Knowl...· 0 citations
A design-oriented framework, an evaluation protocol, and a set of open problems for governed structured-data agents are proposed for retrieval semantics, authorization, intent recognition, entity resolution, evaluation, failure modes, and latency.
The design realization and evaluation of an Automated Summarization Tool (AST) is presented which is a document intelligence platform based on google gemini 2.5 flash that outperforms the strongest fine-tuned transformer baselines (PEGASUS, BART) by ~14 points and is clearly ahead of BERTSUM-ext (a strong transformer baseline), Pointer-Generator Network, TextRank.
K. Kumar, A. Amandeep, Dharmender Kumar et al.· International Journal of Inn...· 0 citations
This work presents DocMaster, a hierarchical structure-aware document analysis system that parses documents into hierarchical document trees preserving original layouts and constructs a structure-aware semantic index that enables accurate document filtering and in-depth analysis.
Ziqi Chen, Yingli Zhou, Fangyuan Zhang et al.· 0 citations
Unstructured exponential growth of NoSQL data is causing an effective challenge for data processing in general, semantic querying, and integration into the traditional analytics pipeline for efficient data processing. Current solutions depend at least in part on rule-based static transformation or partial machine learning models, often lacking any preservation of semantic integrity, schema heterogeneity, or alignment with real-world usage patterns. These constraints severely cripple downstream applications like query optimization, relational mapping, and ML pipeline integration in process. This work proposes a thorough Contextual Structuring Pipeline that automatically converts unstructured NoSQL data into structured, schema-consistent, and optimized relational representations in the queries. The pipeline includes five new learning models designed to contribute to a specific subtask in the transformation. First, CADA-Net (Context-Aware Document Attention Network) uses transformer-based hierarchical encoding to segment raw NoSQL records into semantically meaningful key Value structures. Second, TopoGraph-X constructs an entity-type graph with multi-level topologies from latent document hierarchies through topological learning. Third, HarmoField employs domain-specific BERT embeddings along with the Gaussian mixture model to normalize field representation across heterogeneous sources. Fourth, MetaRel-Frame constructs meta-relational abstractions on logical entity-role patterns to discover relational table blueprints. Finally, QueryStruct transforms the blueprint into a use-optimized dataset through restructuring, indexing, reordering, and compressing its components, taking into account historic query logs. The structural accuracy achieved by this framework is 92.3%, schema normalized is 91.2%, and query latency is decreased by 39.4%. The entire process of structuring unstructured NoSQL data forms a fully automated method, semantically grounded and performance-aware, towards compatibility with SQL-ready, graph analytics, and downstream workflows in machine learning process.
P. Balley, D. Sonekar· Journal of Intelligent Decis...· 0 citations
Large Language Models are deployed in financial applications such as research synthesis and risk analysis, yet their effectiveness is constrained by the limitations of conventional retrieval methods. Existing approaches rely primarily on semantic similarity or token-level matching, which fails in structured domains like finance where relevance depends on precise alignment across entity, temporal and document-type dimensions. This paper proposes the Financial Knowledge Integration Framework (FKIF), a hybrid metadata-aware retrieval system that integrates dense semantic similarity, sparse lexical matching, and structured metadata signals into a unified ranking function. Unlike conventional hybrid retrieval approaches, FKIF treats metadata as a first-class relevance signal rather than an auxiliary feature. Tested on a held-out set of 22 queries drawn from a corpus of 57 SEC filings, FKIF achieves an MRR@10 of 0.98 against TF-IDF’s 0.35, BM25’s 0.17, and dense retrieval’s 0.13. The results demonstrate that metadata-aware retrieval significantly enhances retrieval accuracy and provides a foundation for reliable financial Retrieval-Augmented Generation (RAG) systems.
S. Jambula, Srihari Kumar Pendyala, Rajesh Kumar Butteddi et al.· 2026 International Conferenc...· 0 citations