Structured Prediction for Scalable Spreadsheet Table Understanding: From Cell Types to Table Ranges (Extended Version)
2026-08-17 • Information Retrieval
Information RetrievalDatabasesMachine Learning
AI summaryⓘ
The authors worked on making computers better at understanding spreadsheets by teaching them to identify what kind of data is in each cell and where tables are located in the sheet. They created a two-step system: first, a model classifies cell types using a smart algorithm that looks at features and enforces spatial consistency, then a rule-based method finds the table boundaries. They tested their approach on a new dataset of 737 sheets in multiple languages and found their method is almost as accurate as advanced deep-learning models but uses much less computing power. Their approach also beats some baseline methods and competes well with newer large-language-model systems for finding tables.
SpreadsheetCell-Type ClassificationTable DetectionLightGBMConditional Random Field (CRF)Deterministic AlgorithmStructured PredictionCross-ValidationTransformerLarge Language Models (LLM)
Authors
Antoine Gauquier, Ioana Manolescu, Pierre Senellart
Abstract
Spreadsheets are a primary medium for publishing tabular data, yet automatically extracting structured content from them remains difficult due to heterogeneous layouts, diverse file formats, and inconsistent organizational conventions. We address two core tasks in spreadsheet understanding: Cell-Type Classification (CTC), which assigns roles to cells, and Table Detection (TD), which identifies table bounding boxes within sheets. We propose an efficient two-stage pipeline in which a learned CTC model feeds a deterministic TD algorithm. For CTC, we use a LightGBM classifier over 65 structured features together with a pairwise CRF enforcing spatial consistency across the cell grid. Our TD method extracts table ranges from predicted cell types by a deterministic five-stage procedure. For evaluation, we built and share StatSheets, a multilingual benchmark of 737 manually annotated sheets from 14 public data providers across multiple countries and file formats. Under 5-fold cross-validation, our CRF-LightGBM system achieves a Mean File-Macro F1 score of 0.937 on CTC, within 0.6 percentage points of the GPU-based TUTA Transformer, while requiring substantially fewer computational resources. For TD, our deterministic approach outperforms region-based baselines and remains competitive with recent LLM-based systems such as SpreadsheetLLM. These results demonstrate that combining non-linear structured prediction with deterministic range extraction provides a competitive, scalable, and computationally efficient approach to spreadsheet table understanding.