Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint

TL;DR

This paper introduces a novel two-step table-to-text generation system with human-in-the-loop validation, addressing data sparsity issues in biomedical report generation and demonstrating improved accuracy and adaptability with limited data.

Contribution

The paper presents a new two-step architecture with auto-correction, copy mechanism, and synthetic data augmentation for biomedical table-to-text generation, effective under data sparsity constraints.

Findings

Improved precision in copying tabular values (up to 0.13 increase).

Effective adaptation with only 40% of training data.

Validated outputs through human expert evaluation.

Abstract

Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario.