Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Examining Imbalance Effects on Performance and Demographic Fairness of Clinical Language Models

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read For clinical ICD coding, similarity to the majority class, not group size, predicts performance

desk verdict A useful descriptive study of demographic imbalance in ICD coding whose headline claim rests on a statistically shaky correlation; the central finding needs a redo. read the letter →

arxiv 2412.17803 v2 pith:RL6XWOQ3 submitted 2024-12-23 cs.LG

Precious Jones , Weisi Liu , I-Chan Huang , Xiaolei Huang This is my paper · ORCID
classification cs.LG
keywords dataimbalanceclinicallanguagemodelsICDcodepredictiondemographicfairnesscosinedistanceequalitydifferenceMIMIC-IVlong-tailedlabels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks why clinical language models perform unevenly across demographic and insurance groups when predicting ICD-10 codes from hospital discharge notes. It argues that data imbalance matters, but the dominant driver is not how many patients a subgroup contains; it is how far that subgroup's typical label pattern sits from the rest of the dataset. Using three clinical models on a large intensive-care corpus, the authors show that subgroup F1, precision, recall, and AUC all fall as the cosine distance between the subgroup's average label vector and the overall test set's average label vector rises, while subgroup size shows no significant correlation. They conclude that models overfit majority feature patterns, so fairness work should target feature similarity, and that accuracy is an unreliable metric under long-tailed ICD labels.

What carries the argument

The machine doing the work is a two-step comparison. First, each demographic, age, insurance, and intersectional subgroup is summarized by a single vector: the average, over all patients in the subgroup, of the 100-dimensional ICD-10 label vector. The paper then measures dissimilarity as the cosine distance between that subgroup-average vector and the average vector of the entire test set, producing one number per subgroup. Second, for each performance metric it computes the Pearson correlation coefficient across subgroups between those cosine distances and the subgroup's metric value, testing the null hypothesis of no correlation at a 0.05 significance level. The fairness side uses Equality Difference, the summed absolute deviation of a group's performance from the overall performance, to rank which demographic splits are least fair (age is worst). These two tools—cosine distance between label vectors, and Pearson r—carry the central argument.

What would settle it

Recompute the Pearson correlation between cosine distance and F1-macro on the five non-overlapping age bands alone, dropping the 'All' row; if the negative correlation loses significance, the paper's central claim does not survive.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a correlation result: for the top-100 ICD-10 phenotype prediction task, the F1-macro score of Clinical Longformer across all demographic, age, insurance, and intersectional subgroups is strongly negatively correlated with the cosine distance between the subgroup's globally averaged label vector and the whole test set's averaged label vector (r = -0.973, p = 2.76e-10 for F1-macro on the overall grouping), while no such correlation appears with the subgroup's proportion of the test data. The same pattern holds for micro and weighted F1, precision, recall, and AUC, and it is consistent with the descriptive performance tables for ClinicalBERT and GatorTron. The paper reads this as evidence that data imbalance affects performance and fairness chiefly through feature similarity to the majority class: subgroups whose health-condition signature resembles the bulk of the data are predicted well even if small, and subgroups with dissimilar signatures are predicted poorly even if reasonably represented. It also documents that accuracy behaves oppositely, rising with dissimilarity, which it attributes to the exact-match nature of accuracy on long-tailed multi-label data.

Load-bearing premise

The correlation analysis treats the sixteen subgroup rows as independent measurements even though they are overlapping slices of one test set, with the 'All' row at zero distance and average performance anchoring the regression; if those points are not independent, the reported p-values do not establish the conclusion.

Editorial extensions

If this is right

  • Subgroup performance in ICD-code prediction is explained more by how different a subgroup's label distribution is from the whole dataset than by that subgroup's share of the data.
  • Accuracy should be used with caution in long-tailed clinical multi-label settings, since it can rise for dissimilar subgroups even as recall and F1 fall, masking real underperformance.
  • Fairness audits should report equality differences and correlation with dissimilarity, not just subgroup proportions, to detect the kind of disparity documented here.
  • Clinical Longformer's longer input window gives it an overall performance edge, but the dissimilarity-driven performance gap persists across all three models, suggesting the pattern is not an artifact of a single architecture.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's correlation is computed across overlapping partitions with an 'All' anchor point; a direct robustness check would rerun the correlations on non-overlapping subgroups only, such as the five age bands alone, to see whether the negative slope survives without the anchor.
  • A natural extension the paper does not run: train the same Clinical Longformer on data reweighted to match the majority label distribution and test whether subgroup F1 gaps shrink, which would experimentally separate the dissimilarity effect from the sample-size effect.
  • If the mechanism is genuine, the same cosine-distance analysis should predict performance gaps in other long-tailed clinical tasks, such as mortality or readmission prediction; that is a testable prediction beyond this paper.
  • The positive accuracy–dissimilarity correlation implies that dashboards reporting only accuracy for clinical models can hide severe subgroup underperformance, a practical warning for deployment.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper studies how data imbalance affects the performance and demographic fairness of clinical language models on the ICD-10 phenotype inference task using MIMIC-IV. The authors analyze label and demographic imbalances, evaluate ClinicalBERT, GatorTron, and Clinical Longformer across gender, age, race/ethnicity, and insurance subgroups, and measure fairness with an equality-difference metric. The central claim is that subgroup performance is more strongly associated with the cosine distance between the subgroup's globally averaged label vector and the overall test data than with the subgroup's proportion in the test data, which they interpret as evidence that feature similarity to the majority class is a more critical factor than subgroup size.

Significance. The paper addresses an important and under-studied question: how demographic and social-determinant imbalances affect clinical language model performance and fairness. Its strengths include a large, standard benchmark dataset, evaluation of three established clinical models, a broad set of performance metrics, a released codebase, and a descriptive analysis of intersectional imbalance patterns in Tables I and II. If the main comparative claim were well supported, the study would be a useful reference for the health NLP community. However, the primary statistical evidence for that claim, the Pearson correlation analysis in Section V-C, has load-bearing methodological flaws. The non-independence of the observations, the deterministic anchor of the 'All' row, post-hoc model selection, and the mismatch between 'feature similarity' and label-vector distance all undermine the central conclusion as currently stated. The descriptive findings on performance disparities and fairness remain informative, but the paper's headline claim requires substantially reworked analysis and more careful framing.

major comments (4)
  1. [Section V-C(a), Tables V and VI] The Pearson correlations treat the 16 rows ('All', male, female, five age groups, five race/ethnicity groups, three insurance groups) as independent observations, but these are overlapping partitions of the same test set. The 'All' row is not an independent observation: its cosine distance is 0 by construction and its performance equals the overall mean, so it acts as a deterministic anchor that can manufacture or exaggerate the negative slope. The reported p-values (e.g., F1-macro p=2.76e-10) are therefore not trustworthy, and the contrast with the non-significant proportion correlations is not evidence for the paper's central claim. The analysis should account for the nested and overlapping structure, for example with a mixed-effects model or a cluster bootstrap, and should report results with the 'All' row removed.
  2. [Section V-C, model selection] The correlation analysis is conducted only for Clinical Longformer, which the authors selected after observing that it outperformed the other two models ('As previously determined... we used Clinical Longformer for this correlation analysis'). This post-hoc model selection is not accounted for, and it weakens the inferential value of the correlation results. The central claim would be more credible if the correlations were reported for all three models, or if the selection were justified a priori with an appropriate multiple-comparison adjustment.
  3. [Abstract and Section V-C(a] The conclusion that 'feature similarity to the majority class may be a more critical factor' is not supported by the measurement actually used. The paper computes cosine distance between globally averaged label vectors, which is a label-distribution distance, not a feature similarity measure. In addition, the 'majority class' is never defined: the distance is taken to the overall test data, not to the majority demographic group or a majority label vector. The wording overclaims what the analysis can establish. Either the claims should be limited to label-distribution similarity, or the analysis should be extended to textual feature representations.
  4. [Section V-C, Tables V and VI] Many significance tests are performed across multiple performance metrics and several subgroup partitions without any correction for multiple testing. Since the paper's main conclusion rests on the significance in Table V and the non-significance in Table VI, the alpha=0.05 comparisons should be adjusted or explicitly justified as exploratory. This issue compounds the non-independence problem: even if the rows were independent, the sheer number of tests would make chance findings likely.
minor comments (5)
  1. [Section V-B, Eq. (1)] The equality difference definition uses ED_{g,m} on the left-hand side but sums over all groups g in G on the right-hand side, so the left-hand index is misleading. The notation should clarify that ED is computed per demographic category, not per individual group.
  2. [Table II caption] The caption says each value is the 'mean and margin of error of the confidence interval,' but no confidence level is stated. Please specify the confidence level or report standard deviations/standard errors.
  3. [Section V-C, Pearson correlation formula] The notation around x_i and y is confusing: x_i is described as a vector of performance metrics across groups, while y is a vector of distances or proportions. Please define both variables explicitly and state which quantity corresponds to which axis in Tables V and VI.
  4. [Tables V and VI] The label 'Overall Data' for the first analysis block is misleading, since the rows include all partitions plus the 'All' row. Consider renaming it to something like 'All partitions' or 'All groups combined' to avoid implying it is a single overall test set.
  5. [Conclusion] The conclusion repeats the 'feature similarity' phrasing without acknowledging that the analysis measured label-vector distance rather than text or feature similarity. The terminology should be aligned with the actual method or the method should be extended.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper's central claims are empirical correlation measurements, not derivations from fitted parameters or self-citation chains.

full rationale

This paper is an empirical measurement study rather than a derivation. It trains three clinical language models on MIMIC-IV, measures performance and fairness across demographic subgroups, and computes Pearson correlations between subgroup performance and (a) cosine distance between globally averaged label vectors and (b) subgroup proportion. No parameter is fitted and then renamed as a prediction; the target quantity does not appear in the definition of the predictor; and no uniqueness theorem or ansatz is imported from the authors' prior work. The only self-citations (Refs. [1] and [23]) appear in related-work context and are not load-bearing for the paper's conclusions. Section V-C(a)'s hypothesis that 'feature similarity to the majority class may be a more critical factor' is presented as an inference from observed correlations, and the paper explicitly acknowledges in its limitations that cosine distance is only one possible dissimilarity metric. The inclusion of the 'All' row and the overlapping subgroup partitions in Tables V and VI raise legitimate statistical concerns about independence and the anchoring of the correlation, but those are validity issues, not circularity: the correlation is not equivalent by construction to its own inputs. Similarly, the conceptual distance between label-vector dissimilarity and text-feature similarity is a construct-validity concern, not a circular-reasoning step. Accordingly, no specific circular step can be quoted and exhibited, and the honest finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entities; its central claim rests on a hand-chosen label set, a single hospital dataset, an unvalidated equivalence between label-vector distance and feature similarity, and an assumed leakage-free data split.

free parameters (2)
  • Top-100 ICD-10 code threshold = 100 codes
    The label set is the 100 most frequent ICD-10 codes; this hand-chosen cutoff defines the label imbalance and therefore shapes every performance and fairness result.
  • Minimum note token threshold = 30 tokens
    Discharge summaries shorter than 30 tokens are excluded (Appendix A), altering subgroup sizes and the label distribution without sensitivity analysis.
assumptions (4)
  • domain assumption MIMIC-IV discharge summaries and coded diagnoses are a valid benchmark for studying ICD-10 prediction fairness.
    The entire study is built on one hospital's data from Beth Israel Deaconess Medical Center; the authors acknowledge limited generalizability in Section VI.
  • ad hoc to paper The top-100 frequent ICD-10 codes are a reasonable label set for phenotype inference.
    The task is defined by this selection; results could change with a different code set or code granularity.
  • ad hoc to paper Cosine distance between globally averaged label vectors captures the 'feature similarity' that drives model performance.
    The paper equates label-vector distance with feature similarity without measuring text or learned-feature similarity; this is the key interpretive step for the main claim.
  • domain assumption The 80/10/10 split prevents patient leakage and the test set is representative.
    No patient-level split or random seed is specified; MIMIC-IV contains multiple admissions per patient, so an admission-level split could place the same patient in both training and test sets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Examining Imbalance Effects on Performance and Demographic Fairness of Clinical Language Models." pith.science (2026). https://pith.science/paper/RL6XWOQ3

@misc{pith2026241217803,
  author       = {Pith},
  title        = {Pith review of: Examining Imbalance Effects on Performance and Demographic Fairness of Clinical Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RL6XWOQ3}},
  note         = {Machine review of arXiv:2412.17803}
}
read the original abstract

Data imbalance is a fundamental challenge in applying language models to biomedical applications, particularly in ICD code prediction tasks where label and demographic distributions are uneven. While state-of-the-art language models have been increasingly adopted in biomedical tasks, few studies have systematically examined how data imbalance affects model performance and fairness across demographic groups. This study fills the gap by statistically probing the relationship between data imbalance and model performance in ICD code prediction. We analyze imbalances in a standard benchmark data across gender, age, ethnicity, and social determinants of health by state-of-the-art biomedical language models. By deploying diverse performance metrics and statistical analyses, we explore the influence of data imbalance on performance variations and demographic fairness. Our study shows that data imbalance significantly impacts model performance and fairness, but feature similarity to the majority class may be a more critical factor. We believe this study provides valuable insights for developing more equitable and robust language models in healthcare applications.

Figures

Figures reproduced from arXiv: 2412.17803 by the authors.

Figure 1
Figure 1. Overview of label distribution (ICD-10 codes) by ethnicity group. The codes are arranged in descending order of frequency based on the overall data. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Cosine distances of label vectors between Insurance-Ethnicity. [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Examining and Adapting Time for Multilingual Classification via Mixture of Temporal Experts

    cs.CL 2025-02 conditional novelty 6.0 of 10

    A temporal mixture-of-experts model with cluster-based shift signals improves cross-time multilingual document classification and reveals language-specific temporal performance drops.

Reference graph

Works this paper leans on

37 extracted references · 31 canonical work pages · cited by 1 Pith paper

  1. [1]

    Token imbalance adaptation for radiology report generation,

    Y . Wu, I.-C. Huang, and X. Huang, “Token imbalance adaptation for radiology report generation,” in Proceedings of the Conference on Health, Inference, and Learning , ser. Proceedings of Machine Learning Research, B. J. Mortazavi, T. Sarker, A. Beam, and J. C. Ho, Eds., vol

    work page

  2. [2]

    MIMIC-IV , a freely accessible electronic health record dataset,

    A. E. W. Johnson, L. Bulgarelli, L. Shen, A. Gayles, A. Shammout, S. Horng, T. J. Pollard, S. Hao, B. Moody, B. Gow, L.-w. H. Lehman, L. A. Celi, and R. G. Mark, “MIMIC-IV , a freely accessible electronic health record dataset,” Scientific Data , vol. 10, no. 1, p. 1, 2023. [Online]. Available: https://doi.org/10.1038/s41597-022-01899-x

    doi:10.1038/s41597-022-01899-x 2023

  3. [3]

    Fine-tuned generative llm oversam- pling can improve performance over traditional techniques on multiclass imbalanced text classification,

    N. A. Cloutier and N. Japkowicz, “Fine-tuned generative llm oversam- pling can improve performance over traditional techniques on multiclass imbalanced text classification,” in 2023 IEEE International Conference on Big Data (BigData) . IEEE, 2023, pp. 5181–5186

    work page 2023

  4. [4]

    The shaky foundations of large language models and foundation models for electronic health records,

    M. Wornow, Y . Xu, R. Thapa, B. Patel, E. Steinberg, S. Fleming, M. A. Pfeffer, J. Fries, and N. H. Shah, “The shaky foundations of large language models and foundation models for electronic health records,” npj Digital Medicine , vol. 6, no. 1, p. 135, jul 2023. [Online]. Available: https://www.nature.com/articles/s41746-023-00879-8

    work page 2023

  5. [5]

    Imbalanced learning for hospital readmission prediction using national readmission database,

    S. Wang, M. E. Elkin, and X. Zhu, “Imbalanced learning for hospital readmission prediction using national readmission database,” in 2020 IEEE International Conference on Knowledge Graph (ICKG) . IEEE, 2020, pp. 116–122

    work page 2020

  6. [6]

    A comparative study on deep learning models for text classification of unstructured medical notes with various levels of class imbalance,

    H. Lu, L. Ehwerhemuepha, and C. Rakovski, “A comparative study on deep learning models for text classification of unstructured medical notes with various levels of class imbalance,” BMC medical research methodology, vol. 22, no. 1, p. 181, 2022

    work page 2022

  7. [7]

    Class imbalance in out-of-distribution datasets: Improving the robustness of the textcnn for the classification of rare cancer types,

    K. De Angeli, S. Gao, I. Danciu, E. B. Durbin, X.-C. Wu, A. Stroup, J. Doherty, S. Schwartz, C. Wiggins, M. Damesyn, L. Coyle, L. Penberthy, G. D. Tourassi, and H.-J. Yoon, “Class imbalance in out-of-distribution datasets: Improving the robustness of the textcnn for the classification of rare cancer types,” Journal of Biomedical Informatics, vol. 125, p. ...

    work page 2022

  8. [8]

    A survey of methods for addressing class imbalance in deep-learning based natural language processing,

    S. Henning, W. Beluch, A. Fraser, and A. Friedrich, “A survey of methods for addressing class imbalance in deep-learning based natural language processing,” in Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics , A. Vlachos and I. Augenstein, Eds. Dubrovnik, Croatia: Association for Computational Li...

    work page 2023

Show all 37 references
  1. [9]

    A scoping review on the use of machine learning in research on social determinants of health: Trends and research prospects,

    S. Kino, Y .-T. Hsu, K. Shiba, Y .-S. Chien, C. Mita, I. Kawachi, and A. Daoud, “A scoping review on the use of machine learning in research on social determinants of health: Trends and research prospects,” SSM - Population Health , vol. 15, p. 100836, 2021. [Online]. Availabl...

    2021

  2. [10]

    Peeking into a black box, the fairness and generalizability of a MIMIC-III benchmarking model,

    E. R ¨o¨osli, S. Bozkurt, and T. Hernandez-Boussard, “Peeking into a black box, the fairness and generalizability of a MIMIC-III benchmarking model,” Scientific Data , vol. 9, no. 1, p. 24, jan 2022. [Online]. Available: https://www.nature.com/articles/s41597-021-01110-7

    2022

  3. [11]

    Evaluating the impact of social determinants on health prediction in the intensive care unit,

    M. Y . Yang, G. H. Kwak, T. Pollard, L. A. Celi, and M. Ghassemi, “Evaluating the impact of social determinants on health prediction in the intensive care unit,” in Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society , ser. AIES ’23. New York, NY , USA: Asso...

    2023

  4. [12]

    Mimic-iv-note: Deidentified free-text clinical notes (version 2.2),

    A. Johnson, T. Pollard, S. Horng, L. A. Celi, and R. Mark, “Mimic-iv-note: Deidentified free-text clinical notes (version 2.2),”

  5. [13]

    Publicly available clinical BERT embeddings,

    E. Alsentzer, J. Murphy, W. Boag, W.-H. Weng, D. Jindi, T. Naumann, and M. McDermott, “Publicly available clinical BERT embeddings,” in Proceedings of the 2nd Clinical Natural Language Processing Workshop, A. Rumshisky, K. Roberts, S. Bethard, and T. Naumann, Eds. Minneapolis,...

    2019

  6. [14]

    GatorTron: A Large Clinical Language Model to Unlock Patient Information from Unstructured Electronic Health Records,

    X. Yang, A. Chen, N. PourNejatian, H. C. Shin, K. E. Smith, C. Parisien, C. Compas, C. Martin, M. G. Flores, Y . Zhang, T. Magoc, C. A. Harle, G. Lipori, D. A. Mitchell, W. R. Hogan, E. A. Shenkman, J. Bian, and Y . Wu, “GatorTron: A Large Clinical Language Model to Unlock Pat...

    2022 arXiv

  7. [15]

    Clinical- longformer and clinical-bigbird: Transformers for long clinical se- quences,

    Y . Li, R. M. Wehbe, F. S. Ahmad, H. Wang, and Y . Luo, “Clinical- longformer and clinical-bigbird: Transformers for long clinical se- quences,” 2022

    2022

  8. [16]

    Prediction of icd codes with clinical bert embeddings and text augmentation with label balancing using mimic-iii,

    B. Biseda, G. Desai, H. Lin, and A. Philip, “Prediction of icd codes with clinical bert embeddings and text augmentation with label balancing using mimic-iii,” 2020. [Online]. Available: https://arxiv.org/abs/2008.10492

    2020 arXiv

  9. [17]

    Does the magic of bert apply to medical code assignment? a quantitative study,

    S. Ji, M. H ¨oltt¨a, and P. Marttinen, “Does the magic of bert apply to medical code assignment? a quantitative study,” Computers in biology and medicine, vol. 139, p. 104998, 2021

    2021

  10. [18]

    Leveraging hierarchical category knowledge for data-imbalanced multi-label diagnostic text understanding,

    S.-C. Tsai, T.-Y . Chang, and Y .-N. Chen, “Leveraging hierarchical category knowledge for data-imbalanced multi-label diagnostic text understanding,” in Proceedings of the 10th international workshop on health text mining and information analysis (LOUHI) , 2019, pp. 39–43

    2019

  11. [19]

    Multimodal machine learning for automated icd coding,

    K. Xu, M. Lam, J. Pang, X. Gao, C. Band, P. Mathur, F. Papay, A. K. Khanna, J. B. Cywinski, K. Maheshwari, P. Xie, and E. P. Xing, “Multimodal machine learning for automated icd coding,” in Proceedings of the 4th Machine Learning for Healthcare Conference, ser. Proceedings of ...

  12. [20]

    Learning from undercoded clinical records for automated international classification of diseases (icd) coding,

    Y . Jin, Y . Xiong, D. Shi, Y . Lin, L. He, Y . Zhang, J. M. Plasek, L. Zhou, D. W. Bates, and C. Tang, “Learning from undercoded clinical records for automated international classification of diseases (icd) coding,” Journal of the American Medical Informatics Association, vol...

    2023

  13. [21]

    Transformers and large language models in healthcare: A review,

    S. Nerella, S. Bandyopadhyay, J. Zhang, M. Contreras, S. Siegel, A. Bu- min, B. Silva, J. Sena, B. Shickel, A. Bihorac et al., “Transformers and large language models in healthcare: A review,” Artificial Intelligence in Medicine, p. 102900, 2024

    2024

  14. [22]

    The potential of chatgpt as a self- diagnostic tool in common orthopedic diseases: exploratory study,

    T. Kuroiwa, A. Sarcon, T. Ibara, E. Yamada, A. Yamamoto, K. Tsukamoto, and K. Fujita, “The potential of chatgpt as a self- diagnostic tool in common orthopedic diseases: exploratory study,” Journal of Medical Internet Research , vol. 25, p. e47621, 2023

    2023

  15. [23]

    Time matters: Examine temporal effects on biomedical language models,

    W. Liu, Z. He, and X. Huang, “Time matters: Examine temporal effects on biomedical language models,” 2024. [Online]. Available: https://arxiv.org/abs/2407.17638

    2024 arXiv

  16. [24]

    A multimodal transformer: Fusing clinical notes with structured ehr data for interpretable in-hospital mortality prediction,

    W. Lyu, X. Dong, R. Wong, S. Zheng, K. Abell-Hart, F. Wang, and C. Chen, “A multimodal transformer: Fusing clinical notes with structured ehr data for interpretable in-hospital mortality prediction,” in AMIA Annual Symposium Proceedings , vol. 2022. American Medical Informatic...

    2022

  17. [25]

    Dalk: Dynamic co-augmentation of llms and kg to answer alzheimer’s disease questions with scientific literature,

    D. Li, S. Yang, Z. Tan, J. Y . Baik, S. Yun, J. Lee, A. Chacko, B. Hou, D. Duong-Tran, Y . Dinget al., “Dalk: Dynamic co-augmentation of llms and kg to answer alzheimer’s disease questions with scientific literature,” arXiv preprint arXiv:2405.04819 , 2024

    2024 arXiv

  18. [26]

    Thought graph: Generating thought process for biological reasoning,

    C.-Y . Hsu, K. Cox, J. Xu, Z. Tan, T. Zhai, M. Hu, D. Pratt, T. Chen, Z. Hu, and Y . Ding, “Thought graph: Generating thought process for biological reasoning,” in Companion Proceedings of the ACM on Web Conference 2024, 2024, pp. 537–540

    2024

  19. [27]

    Improving imbalanced learning by pre-finetuning with data augmentation,

    Y . Shi, T. ValizadehAslani, J. Wang, P. Ren, Y . Zhang, M. Hu, L. Zhao, and H. Liang, “Improving imbalanced learning by pre-finetuning with data augmentation,” in Proceedings of the Fourth International Work- shop on Learning with Imbalanced Domains: Theory and Applications ,...

    2022

  20. [28]

    Physiobank, physiotoolkit, and physionet: Components of a new research resource for complex physiologic signals,

    A. L. Goldberger, L. A. Amaral, L. Glass, J. M. Hausdorff, P. C. Ivanov, R. G. Mark, J. E. Mietus, G. B. Moody, C.-K. Peng, and H. E. Stanley, “Physiobank, physiotoolkit, and physionet: Components of a new research resource for complex physiologic signals,” Circulation, vol. 1...

    2000

  21. [29]

    Gender determinants of cardiovascular risk factors and diseases,

    G. Mercuro, M. Deidda, A. Piras, C. C. Dessalvi, S. Maffei, and G. M. Rosano, “Gender determinants of cardiovascular risk factors and diseases,” Journal of Cardiovascular Medicine , vol. 11, no. 3, pp. 207– 220, 2010

    2010

  22. [30]

    Autoimmune disease and gender: plausible mechanisms for the female predominance of autoimmunity,

    O. L. Quintero, M. J. Amador-Patarroyo, G. Montoya-Ortiz, A. Rojas- Villarraga, and J.-M. Anaya, “Autoimmune disease and gender: plausible mechanisms for the female predominance of autoimmunity,” Journal of autoimmunity, vol. 38, no. 2-3, pp. J109–J119, 2012

    2012

  23. [31]

    A comparative study of pretrained language models for long clinical text,

    Y . Li, R. M. Wehbe, F. S. Ahmad, H. Wang, and Y . Luo, “A comparative study of pretrained language models for long clinical text,” Journal of the American Medical Informatics Association , vol. 30, no. 2, pp. 340– 347, 2023

    2023

  24. [32]

    Measuring and mitigating unintended bias in text classification,

    L. Dixon, J. Li, J. Sorensen, N. Thain, and L. Vasserman, “Measuring and mitigating unintended bias in text classification,” in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society , 2018, pp. 67–73

    2018

  25. [33]

    Icd-mappings,

    S. Gonc ¸alves and G. Ehrensperger, “Icd-mappings,” 2023. [Online]. Available: https://github.com/snovaisg/ICD-Mappings

    2023

  26. [34]

    Transformers: State- of-the-art natural language processing,

    T. Wolf, L. Debut, V . Sanh, J. Chaumond, C. Delangue, A. Moi, P. Cistac, T. Rault, R. Louf, M. Funtowicz, J. Davison, S. Shleifer, P. von Platen, C. Ma, Y . Jernite, J. Plu, C. Xu, T. Le Scao, S. Gugger, M. Drame, Q. Lhoest, and A. Rush, “Transformers: State- of-the-art natur...

    2020

  27. [106]

    PMLR, 09–10 Aug 2019, pp. 197–215. [Online]. Available: https://proceedings.mlr.press/v106/xu19a.html

    2019

  28. [209]

    PMLR, 22 Jun–24 Jun 2023, pp. 72–85. [Online]. Available: https://proceedings.mlr.press/v209/wu23a.html

    2023

  29. [2023]

    Available: https://doi.org/10.13026/1n74-ne17

    [Online]. Available: https://doi.org/10.13026/1n74-ne17

    doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.