Joe D. Menke

Shufan Ming, Joe D. Menke, Neil R. Smalheiser et al.

Accurately and consistently indexing biomedical literature by publication type and study design is essential for supporting evidence synthesis and knowledge discovery. Prior work on automated publication type and study design indexing has primarily focused on expanding label coverage, enriching feature representations, and improving in-domain accuracy, with evaluation typically conducted on data drawn from the same distribution as training. Although pretrained biomedical language models achieve strong performance under these settings, models optimized for in-domain accuracy may rely on superficial lexical or dataset-specific cues, resulting in reduced robustness under distributional shift. In this study, we introduce an evaluation framework based on controlled semantic perturbations to assess the robustness of a publication type classifier and investigate robustness-oriented training strategies that combine entity masking and domain-adversarial training to mitigate reliance on spurious topical correlations. Our results show that the commonly observed trade-off between robustness and in-domain accuracy can be mitigated when robustness objectives are designed to selectively suppress non-task-defining features while preserving salient methodological signals. We find that these improvements arise from two complementary mechanisms: (1) increased reliance on explicit methodological cues when such cues are present in the input, and (2) reduced reliance on spurious domain-specific topical features. These findings highlight the importance of feature-level robustness analysis for publication type and study design classification and suggest that refining masking and adversarial objectives to more selectively suppress topical information may further improve robustness. Data, code, and models are available at: https://github.com/ScienceNLP-Lab/MultiTagger-v2/tree/main/ICHI

Neil R. Smalheiser, Joe D. Menke, Arthur W. Holt et al.

Objectives. Major research and implementation efforts have been devoted to indexing articles according to the major topics discussed, but much less effort to indexing their publication types and study designs (collectively, PTs). In this Perspective, we discuss how indexing PTs differs from topical MeSH indexing and requires a different approach. Materials and Methods. Rather than focus on the technical aspects of machine learning-based indexing models, we emphasize the goals and purposes for which biomedical articles are indexed, and the surprisingly thorny question of how indexing systems should be evaluated. Results. Topical Medical Subject Heading (MeSH) terms are assigned to articles that cover the major topics discussed; when more than one term is applicable, only the most specific term is assigned. In contrast, PTs are assigned to articles that have a given structure or use a particular design. To meet the needs of end-users, particularly groups involved in evidence syntheses, PT indexing needs to be comprehensive and employ probabilistic goodness-of-fit prediction scores. Whereas existing NLM hierarchies place publication types and study design-related terms on separate trees from each other, we have created a unified hierarchy that permits more appropriate retrieval via automatic expansion. Discussion. Automated PT indexing systems should allow users to input article records or full-text PDFs and receive scores in real time. This will offer consistent indexing across bibliographic databases, as well as preprints and unpublished manuscripts. Conclusions. Automated PT indexing systems, properly designed and implemented, hold the promise of greatly improving the retrieval of biomedical articles, saving substantial effort when writing evidence syntheses and benefiting other users as well.