analiza aiqbench
MedStruct-S introduces a critical benchmark for evaluating AI models on semi-structured information extraction from noisy OCR clinical data, specifically addressing the challenges of unknown keys and heterogeneous representations. The study reveals that encoder-only architectures often outperform larger decoder-only models in key-conditioned QA tasks when model scale is not controlled, offering practical guidance for selecting efficient solutions in real-world medical history reconstruction.
MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
arXiv:2605.03103v2 Announce Type: replace-cross Abstract: Semi-structured information extraction (IE) from OCR-derived clinical reports is crucial for efficiently reconstructing patients' longitudinal medical histories. In practice, this scenario commonly involves three tasks: (i) field-header (key) discovery, (ii) key-conditioned question answering (QA), and (iii) end-to-end key-value pair extraction. However, existing evaluations often under-model two factors: heterogeneous and incompletely known key representations, and OCR-induced noise. This makes it difficult to assess model robustness in real-world settings. We present MedStruct-S, a benchmark specifically designed to evaluate these tasks under unknown keys and OCR noise. MedStruct-S contains 3,582 annotated real-world clinical report pages. Using MedStruct-S, we benchmark two representative paradigms: encoder-only sequence labeling with post-processing and decoder-only structured generation, covering four encoder-only and five decoder-only models spanning 0.11B to 103B parameters. Our results show that encoder-only models achieve the best performance for non-null-value key-conditioned QA despite being substantially smaller than decoder-only models. When comparing models of similar order of magnitude, encoder-only models still perform better overall. Without controlling for model scale, fine-tuned decoder-only models deliver the strongest overall results. These findings show that the benchmark provides a reliable and practical basis for selecting and comparing models across different semi-structured IE settings.
Abstract:Semi-structured information extraction (IE) from OCR-derived clinical reports is crucial for efficiently reconstructing patients' longitudinal medical histories. In practice, this scenario commonly involves three tasks: (i) field-header (key) discovery, (ii) key-conditioned question answering (QA), and (iii) end-to-end key-value pair extraction. However, existing evaluations often under-model two factors: heterogeneous and incompletely known key representations, and OCR-induced noise. This makes it difficult to assess model robustness in real-world settings. We present MedStruct-S, a benchmark specifically designed to evaluate these tasks under unknown keys and OCR noise. MedStruct-S contains 3,582 annotated real-world clinical report pages. Using MedStruct-S, we benchmark two representative paradigms: encoder-only sequence labeling with post-processing and decoder-only structured generation, covering four encoder-only and five decoder-only models spanning 0.11B to 103B parameters. Our results show that encoder-only models achieve the best performance for non-null-value key-conditioned QA despite being substantially smaller than decoder-only models. When comparing models of similar order of magnitude, encoder-only models still perform better overall. Without controlling for model scale, fine-tuned decoder-only models deliver the strongest overall results. These findings show that the benchmark provides a reliable and practical basis for selecting and comparing models across different semi-structured IE settings.
11 pages, 5 figures. Accepted by KSEM 2026. This is the author's preprint version; the final authenticated version will be available in the Springer LNCS/LNAI proceedings
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
I.2.7; H.3.1; J.3
arXiv:2605.03103 [cs.CL]
(or arXiv:2605.03103v2 [cs.CL] for this version)
Text integral sub licența deschisă a publicației · 237 cuvinte
