Pith. sign in

REVIEW 3 major objections 5 minor 37 references

The Anatomy of Evidence: An Investigation Into Explainable ICD Coding

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Proximity-aware match measures show explainable ICD models align with human evidence in about 80% of test cases.

desk verdict A useful, honest evaluation paper with new match measures and dataset analyses; the 80% headline is recall-oriented, and the lack of a precision baseline is the main gap. read the letter →

arxiv 2507.01802 v1 pith:P76DQHNT submitted 2025-07-02 cs.CL cs.LG

Katharina Beckh , Elisa Studeny , Sujan Sai Gannamaneni , Dario Antweiler , Stefan Rüping This is my paper
classification cs.CLcs.LG
keywords ICDcodingexplainabilityevidenceextractionMDACEdatasetmatchmeasuresplausibilityevaluationmedicalfeatureattribution
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 argues that evaluating explainable medical coding with proximity-aware match measures—empty, exact, proximate, partial, and no match—reveals a practically useful alignment between extracted evidence and human-annotated ground truth. Using the MDACE dataset, it shows that a supervised model trained on evidence annotations identifies at least one correct token in roughly 80% of test cases, and that supervised training produces more exact matches and fewer empty or no-match cases than an unsupervised baseline. This matters because medical coding needs transparency for coders and regulators, and the paper provides an evaluation protocol that captures whether an explanation guides a human to the right context, which is more meaningful than F1 or intersection-over-union scores. The paper also analyzes where evidence appears in documents, how much it overlaps with code descriptions, and how sufficient versus complete annotations differ, yielding recommendations for developing and evaluating explainable ICD coding systems.

What carries the argument

The central object is a taxonomy of five match measures that compare model-extracted evidence spans against ground truth spans: Empty (no evidence predicted), Exact (the sets of spans are equal), Proximate (every ground truth span has at least one matching token and all unmatched tokens fall within a context window of k=10 tokens), Partial (at least one ground truth span lacks a match or some model tokens fall outside the context window), and No match (no token overlap). This taxonomy carries the evaluation argument because it treats evidence as useful when it guides attention to the right context window, rather than requiring exact span equality as F1 and IOU do. The supporting machinery is the MDACE dataset with its sufficient and complete code-evidence annotations, the PLM-ICD-based architecture modified with standard cross-attention, and the AttInGrad feature attribution method (attention weights multiplied by input-times-gradient scores) with a threshold chosen on the validation set.

What would settle it

Re-annotate a random sample of the 61 test discharge summaries with an independent team of professional coders under the same guidelines, then compute inter-annotator agreement between the new annotations and the MDACE spans; if that agreement is no higher than the supervised model's match rate on the same documents, the ground truth is too noisy to support the paper's comparative claims.

Watch

Extended reading notes

Core claim

The central claim is that current explainable ICD coding models, when evaluated with proximity-aware match categories instead of F1/IOU, align substantially with human-annotated evidence: the best supervised model achieves some form of match (exact, proximate, or partial) in about 80% of test cases, with relatively few empty or no-match predictions. The supervised approach, which uses annotated evidence in its training objective, yields more exact matches and fewer empty and no-match cases than an unsupervised baseline that only regularizes gradients. The paper also finds that human evidence is not systematically located at the document start, that many codes have little overlap between evidence and their official descriptions, and that the 'sufficient' (Inpatient) annotation scheme is not a subset of the 'complete' (Profee) scheme in the common documents. These findings support a recommendation for investment in evidence annotation when exact matches are desired, and for reporting false positives in future evaluation.

Load-bearing premise

The evaluation assumes that the MDACE human-annotated evidence spans are a complete and reliable gold standard, and that the 61-document test set with 586 evidence spans, together with one selected model seed per training strategy, is enough to compare supervised versus unsupervised evidence extraction.

Editorial extensions

If this is right

  • If the central claim holds, supervised training with evidence annotations should be the recommended strategy when deployment requires exact or near-exact evidence matches, and investing in evidence annotation is a concrete way to improve that outcome.
  • The proximity-aware match measures constitute a ready-to-use evaluation protocol for explainable medical coding that is more interpretable to end users than F1 or IOU, and can be adopted as a default reporting standard alongside existing metrics.
  • The finding that many codes have high overlap between evidence and code descriptions suggests that a subset of the ICD coding task can be solved efficiently with rule-based or hybrid approaches, freeing deep learning for the harder, low-overlap codes.
  • The observed relationship between short model evidence and lower recall indicates that explanation length can serve as a lightweight warning signal for likely classification errors in deployed systems.
  • The low probabilities associated with empty evidence predictions and the absence of a probability signal for exact matches give a concrete direction for future work on model calibration and its interaction with explanation quality.

Reading between the lines

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

  • A natural extension of the match taxonomy would be to apply it to other evidence-extraction tasks in clinical NLP, such as procedure coding or coding in languages other than English, where the same proximity semantics could guide evaluation design.
  • The 80% overlap figure is dominated by partial matches; if a stricter bar requiring exact or proximate matches were used, the practical usefulness of current models would appear considerably lower, which is worth stating explicitly when comparing systems.
  • The paper's observation that models extract drugs and co-morbid conditions as evidence for diagnoses suggests that feature attribution methods reflect correlational rather than causal structure; regulatory acceptance of such explanations may require additional causal or counterfactual validation.
  • The finding that sufficient evidence is not a subset of complete evidence, while the authors attribute it to different coding rules, may also indicate that the annotation guidelines themselves need formalization before a single gold standard can be assumed for benchmark comparisons.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper analyzes the MDACE dataset and evaluates current explainable ICD coding models. It studies where human-annotated evidence is located, how much evidence overlaps with code descriptions, and whether sufficient evidence is a subset of complete evidence. For model evaluation, it introduces match measures (empty, exact, proximate, partial, no match) and reports that the best supervised model identifies at least one correct token in roughly 80% of test cases, with supervised training producing more exact matches and fewer empty or non-matches than an unsupervised input-gradient-regularization model. The paper closes with recommendations for practitioners and researchers, including the reporting of false positives.

Significance. If substantiated, the paper offers a useful application-oriented evaluation protocol for evidence extraction in medical coding and a helpful characterization of the MDACE dataset. The strengths are the public code, the transparent match counts, the explicit split and seed details, and the qualitative error analysis. However, the headline overlap figure and the supervised-versus-unsupervised comparison depend on permissive match definitions and on a single selected seed per training strategy, and the paper does not yet provide a baseline or precision analysis that would show the match measures capture evidence quality better than F1/IOU. The contribution is therefore potentially valuable to the community, but the central empirical claims need additional support before the recommendations can be taken as established.

major comments (3)
  1. [§4.3, §4.1, §4.5] The headline 80% overlap result counts exact, proximate, and partial matches as success, but the proximate and partial definitions are very permissive. Proximate requires each ground-truth sequence to share at least one token, with unmatched tokens inside a context window of k=10; partial is triggered whenever at least one ground-truth sequence is unmatched or any machine token falls outside the window. No constraint limits the total number of machine-extracted tokens or false positives, and §4.5 itself notes that models "tend to extract more evidence than humans annotated, often echoing the same or similar information at different positions." Without a trivial high-recall baseline (e.g., top-k attribution tokens or all tokens above a low threshold) and without precision or false-positive counts per match type, the 80% figure and the claim that the proposed measures are more meaningful than F1/IOU are not established; the results may reflect extraction breadth rather than evidence quality.
  2. [§4.1, §5.1, Appendix C.1] The comparison between supervised and unsupervised training strategies is based on one seed per strategy, selected according to explanation metrics. Because ten model seeds are available for each strategy and Figure 8 shows nontrivial variation across seeds in F1, precision, and recall, the observed differences in exact, empty, and no-match counts cannot currently be attributed to the training strategy rather than to the selected seed or to selection effects. The recommendation that investing in evidence annotation improves exact matches would require an aggregate comparison over seeds with appropriate error bars or a paired statistical test.
  3. [§4.3] The semantic relabeling of no-match cases was performed by two authors with "basic medical knowledge," without inter-annotator agreement and without considering the surrounding context; the authors themselves state that expert assessments are necessary. The reported 46% and 53% "semantically meaningful" rates should therefore be treated as exploratory and should not carry quantitative weight in the narrative that model evidence is closer to human evidence than token-level matching suggests.
minor comments (5)
  1. [§3.3] "we selected the 52 document with common hospital admission ID" appears to be a typo for "52 documents."
  2. [Figure 6] The figure uses the abbreviation "prox." while the text defines "proximate"; please use the same term consistently.
  3. [Figure 5] The figure caption does not identify the blue/orange color scheme that the text refers to; a legend or explicit caption labels would improve readability.
  4. [§4.4] The 74% agreement figure is computed for two selected seeds; the text should state explicitly that this is a seed-specific observation rather than a property of the training strategies in general.
  5. [§5, Limitations] The fixed context window k=10 is acknowledged in the limitations, but a short sensitivity analysis varying k would substantially strengthen the proximate-match definition.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the evaluation uses external human-annotated evidence and independently defined match measures; disclosed threshold dependence is a limitation, not a circular step.

full rationale

The paper's central claims—that human-annotated evidence aligns with code descriptions to a certain degree, and that supervised evidence extraction achieves high overlap with ground truth under proximity-aware match measures—are evaluated against the externally produced MDACE dataset (Cheng et al., 2023), not against quantities defined by the paper's own conclusions. The proposed match measures (empty, exact, proximate, partial, no match) are defined independently of the models being evaluated and are computed from token overlap between model attributions and human-annotated spans; this is a standard external plausibility evaluation rather than a self-definitional exercise. The 80% overlap figure in Section 4.3 is an empirical count on the held-out test set, not a fitted parameter or an identity derived from the measure definitions. The paper's self-citations (Antweiler et al., 2023; Klug et al., 2024; Müller et al., 2023) and the use of Edin et al. (2024) models are contextual or comparative and do not carry the weight of the main evidence. The one disclosed dependence—that the unsupervised model's attribution threshold is set using annotated evidence (Section 5.1: 'The unsupervised training strategy with input gradient regularization proved similarly successful as the supervised approach, the annotated evidence was still used to compute a feature attribution threshold')—is an honest limitation that weakens the interpretation of the unsupervised comparison, but it does not make the match-rate results circular, since the threshold is a calibration choice and the match rates are still measured against external ground truth. Likewise, the permissiveness of proximate/partial matches is a methodological concern about whether the measure captures usable evidence quality, not a circularity in the derivation. No step was found where a predicted quantity is equivalent by construction to an input or where a load-bearing premise reduces to a self-citation chain.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The paper's central claims rest on external human annotations (MDACE), on feature attribution scores as evidence, and on two hand-picked model seeds. No novel theoretical entities are introduced; the free parameters k, the prediction threshold, and the attribution threshold are all user- or validation-set choices that the match results are sensitive to.

free parameters (3)
  • Proximate match context window k = 10
    Chosen by authors in Section 4.1; all proximate and partial match counts, and the 80% overlap result, depend on this window size.
  • Code prediction probability threshold = 0.5
    Section 4.4 treats all codes with output probability at least 0.5 as predicted; this determines the true positive set used in the match analysis.
  • AttInGrad attribution decision threshold = not reported
    Section 4.1 sets this threshold on the validation set using annotated evidence; it controls which tokens are extracted as model evidence, and the supervised versus unsupervised comparison depends on it.
assumptions (3)
  • domain assumption Human-annotated evidence spans in MDACE are a valid gold standard for explanation plausibility.
    Sections 3 and 4.3 treat all human spans as ground truth. The paper itself notes annotation dependencies and underspecification in Section 3.4, so this assumption is load-bearing but only partially validated.
  • domain assumption Feature attribution scores from AttInGrad are meaningful indicators of model evidence.
    Section 4.1 adopts AttInGrad because prior work found it best on faithfulness and plausibility, but the paper does not independently verify that attribution scores reflect the model's actual reasoning.
  • ad hoc to paper A single seed per training strategy is representative for comparing supervised versus unsupervised evidence extraction.
    Section 4.1 selects one supervised and one unsupervised seed based on explanation F1. The RQ5 and RQ6 conclusions generalize from only these two seeds, with no variance analysis across the 50 available model weights.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Anatomy of Evidence: An Investigation Into Explainable ICD Coding." pith.science (2026). https://pith.science/paper/P76DQHNT

@misc{pith2026250701802,
  author       = {Pith},
  title        = {Pith review of: The Anatomy of Evidence: An Investigation Into Explainable ICD Coding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P76DQHNT}},
  note         = {Machine review of arXiv:2507.01802}
}
read the original abstract

Automatic medical coding has the potential to ease documentation and billing processes. For this task, transparency plays an important role for medical coders and regulatory bodies, which can be achieved using explainability methods. However, the evaluation of these approaches has been mostly limited to short text and binary settings due to a scarcity of annotated data. Recent efforts by Cheng et al. (2023) have introduced the MDACE dataset, which provides a valuable resource containing code evidence in clinical records. In this work, we conduct an in-depth analysis of the MDACE dataset and perform plausibility evaluation of current explainable medical coding systems from an applied perspective. With this, we contribute to a deeper understanding of automatic medical coding and evidence extraction. Our findings reveal that ground truth evidence aligns with code descriptions to a certain degree. An investigation into state-of-the-art approaches shows a high overlap with ground truth evidence. We propose match measures and highlight success and failure cases. Based on our findings, we provide recommendations for developing and evaluating explainable medical coding systems.

Figures

Figures reproduced from arXiv: 2507.01802 by the authors.

Figure 1
Figure 1. Discharge summary with ICD codes and re [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Average number of evidence spans for each [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Distribution of human-annotated evidence [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Distribution of ICD codes according to evi [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 6
Figure 6. Figure 6: Comparison of the supervised and unsuper [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Alignment of match types between supervised [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: F1 scores, precision and recall in regards to [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

37 extracted references · 14 canonical work pages

  1. [1]

    Dario Antweiler, Florian Gallusser, and Georg Fuchs. 2023. https://doi.org/10.1109/vahc60858.2023.00010 Multi-task transformer visualization to build trust for clinical outcome prediction . In 2023 Workshop on Visual Analytics in Healthcare (VAHC), volume 34, page 21–26. IEEE

    work page arXiv 2023

  2. [2]

    Pepa Atanasova, Jakob Grue Simonsen, Christina Lioma, and Isabelle Augenstein. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.263 A diagnostic study of explainability techniques for text classification . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 3256--3274, Online. Association for Computati...

    doi:10.18653/v1/2020.emnlp-main.263 2020

  3. [3]

    Biplob Biswas, Thai-Hoang Pham, and Ping Zhang. 2021. https://doi.org/10.1007/978-3-030-77211-6_56 TransICD: Transformer Based Code-Wise Attention Model for Explainable ICD Coding , page 469–478. Springer International Publishing

    work page doi:10.1007/978-3-030-77211-6_56 2021

  4. [4]

    Hua Cheng, Rana Jafari, April Russell, Russell Klopfer, Edmond Lu, Benjamin Striner, and Matthew Gormley. 2023. https://doi.org/10.18653/v1/2023.acl-long.416 MDACE : MIMIC documents annotated with code evidence . In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7534--7550, Toronto, C...

    doi:10.18653/v1/2023.acl-long.416 2023

  5. [5]

    Xiang Dai, Ilias Chalkidis, Sune Darkner, and Desmond Elliott. 2022. https://doi.org/10.18653/v1/2022.findings-emnlp.534 Revisiting transformer-based models for long document classification . In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 7212--7230, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics

    doi:10.18653/v1/2022.findings-emnlp.534 2022

  6. [6]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. https://doi.org/10.18653/v1/N19-1423 BERT : Pre-training of deep bidirectional transformers for language understanding . In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long a...

    doi:10.18653/v1/n19-1423 2019

  7. [7]

    Jay DeYoung, Sarthak Jain, Nazneen Fatema Rajani, Eric Lehman, Caiming Xiong, Richard Socher, and Byron C. Wallace. 2020. https://doi.org/10.18653/v1/2020.acl-main.408 ERASER : A benchmark to evaluate rationalized NLP models . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4443--4458, Online. Association ...

    doi:10.18653/v1/2020.acl-main.408 2020

  8. [8]

    Havtorn, Lasse Borgholt, Maria Maistro, Tuukka Ruotsalo, and Lars Maal e

    Joakim Edin, Alexander Junge, Jakob D. Havtorn, Lasse Borgholt, Maria Maistro, Tuukka Ruotsalo, and Lars Maal e. 2023. https://doi.org/10.1145/3539618.3591918 Automated medical coding on MIMIC-III and MIMIC-IV : A critical review and replicability study . In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Informat...

    arXiv 2023

Show all 37 references
  1. [9]

    Joakim Edin, Maria Maistro, Lars Maal e, Lasse Borgholt, Jakob Drachmann Havtorn, and Tuukka Ruotsalo. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.280 An unsupervised approach to achieve supervised-level explainability in healthcare records . In Proceedings of the 2024 C...

    2024 doi

  2. [10]

    Ary L Goldberger, Luis AN Amaral, Leon Glass, Jeffrey M Hausdorff, Plamen Ch Ivanov, Roger G Mark, Joseph E Mietus, George B Moody, Chung-Kang Peng, and H Eugene Stanley. 2000. Physiobank, physiotoolkit, and physionet: components of a new research resource for complex physiolo...

    2000

  3. [11]

    Erick Mendez Guzman, Viktor Schlegel, and Riza Batista - Navarro. 2024. https://doi.org/10.3389/FRAI.2024.1363531 From outputs to insights: a survey of rationalization approaches for explainable text classification . Frontiers Artif. Intell., 7

    2024

  4. [12]

    Gaspard Harerimana, Jong Wook Kim, and Beakcheol Jang. 2021. https://doi.org/10.1016/j.jbi.2021.103778 A deep attention model to forecast the length of stay and the in-hospital mortality right on admission from icd codes and demographic data . Journal of Biomedical Informatics...

    2021

  5. [13]

    Matthew Honnibal, Ines Montani, Sofie Van Landeghem, and Adriane Boyd. 2020. https://doi.org/10.5281/zenodo.1212303 spaCy: Industrial-strength Natural Language Processing in Python

    2020 doi

  6. [14]

    Wen-hui Hou, Xiao-kang Wang, Ya-nan Wang, Jian-qiang Wang, and Fei Xiao. 2024. https://doi.org/10.1016/j.eswa.2024.123519 Modelling long medical documents and code associations for explainable automatic icd coding . Expert Systems with Applications, 249:123519

    2024

  7. [15]

    Chao-Wei Huang, Shang-Chi Tsai, and Yun-Nung Chen. 2022. https://doi.org/10.18653/v1/2022.clinicalnlp-1.2 PLM - ICD : Automatic ICD coding with pretrained language models . In Proceedings of the 4th Clinical Natural Language Processing Workshop, pages 10--20, Seattle, WA. Asso...

    2022 doi

  8. [16]

    Adam Ivankay, Mattia Rigotti, and Pascal Frossard. 2023. https://doi.org/10.18653/v1/2023.acl-long.644 DARE : Towards robust text explanations in biomedical and healthcare applications . In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics...

    2023 doi

  9. [17]

    Alon Jacovi and Yoav Goldberg. 2020. https://doi.org/10.18653/v1/2020.acl-main.386 Towards faithfully interpretable NLP systems: How should we define and evaluate faithfulness? In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 41...

    2020 doi

  10. [18]

    Shaoxiong Ji, Xiaobo Li, Wei Sun, Hang Dong, Ara Taalas, Yijia Zhang, Honghan Wu, Esa Pitk \"a nen, and Pekka Marttinen. 2024. A unified review of deep learning for automated medical coding. ACM Computing Surveys, 56(12):1--41

    2024

  11. [19]

    Alistair EW Johnson, Tom J Pollard, Lu Shen, Li-wei H Lehman, Mengling Feng, Mohammad Ghassemi, Benjamin Moody, Peter Szolovits, Leo Anthony Celi, and Roger G Mark. 2016. MIMIC-III , a freely accessible critical care database. Scientific Data, 3(1):1--9

    2016

  12. [20]

    Byung-Hak Kim, Zhongfen Deng, Philip Yu, and Varun Ganapathi. 2022. https://doi.org/10.18653/v1/2022.louhi-1.3 Can current explainability help provide references in clinical notes to support humans annotate medical codes? In Proceedings of the 13th International Workshop on He...

    2022 doi

  13. [21]

    Katrin Klug, Katharina Beckh, Dario Antweiler, Nilesh Chakraborty, Giulia Baldini, Katharina Laue, Ren \' e Hosch, Felix Nensa, Martin Schuler, and Sven Giesselbach. 2024. https://doi.org/10.1186/S12911-024-02641-W From admission to discharge: a systematic review of clinical n...

    2024 doi

  14. [22]

    Patrick Lewis, Myle Ott, Jingfei Du, and Veselin Stoyanov. 2020. https://doi.org/10.18653/v1/2020.clinicalnlp-1.17 Pretrained language models for biomedical and clinical tasks: Understanding and extending the state-of-the-art . In Proceedings of the 3rd Clinical Natural Langua...

    2020 doi

  15. [23]

    Nguyen, Vicki Bennett, and Louisa Jorm

    Leibo Liu, \' O scar P \' e rez Concha, Anthony N. Nguyen, Vicki Bennett, and Louisa Jorm. 2022. https://doi.org/10.1016/J.JBI.2022.104161 Hierarchical label-wise attention transformer model for explainable ICD coding . Journal of Biomedical Informatics, 133:104161

    2022

  16. [24]

    Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. https://arxiv.org/abs/1907.11692 RoBERTa : A robustly optimized BERT pretraining approach . Preprint, arXiv:1907.11692

    2019 arXiv

  17. [25]

    Scott M Lundberg and Su-In Lee. 2017. https://proceedings.neurips.cc/paper_files/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf A unified approach to interpreting model predictions . In Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc

    2017

  18. [26]

    Junyu Luo, Xiaochen Wang, Jiaqi Wang, Aofei Chang, Yaqing Wang, and Fenglong Ma. 2024. https://aclanthology.org/2024.lrec-main.355 C o R elation: Boosting automatic ICD coding through contextualized code relation learning . In Proceedings of the 2024 Joint International Confer...

    2024

  19. [27]

    James Mullenbach, Sarah Wiegreffe, Jon Duke, Jimeng Sun, and Jacob Eisenstein. 2018. https://doi.org/10.18653/v1/N18-1100 Explainable prediction of medical codes from clinical text . In Proceedings of the 2018 Conference of the North A merican Chapter of the Association for Co...

    2018 doi

  20. [28]

    Sebastian M \" u ller, Vanessa Toborek, Katharina Beckh, Matthias Jakobs, Christian Bauckhage, and Pascal Welke. 2023. https://doi.org/10.1007/978-3-031-43418-1\_28 An empirical evaluation of the Rashomon Effect in explainable machine learning . In Machine Learning and Knowled...

    2023 doi

  21. [29]

    o rg Schl \

    Meike Nauta, Jan Trienes, Shreyasi Pathak, Elisa Nguyen, Michelle Peters, Yasmin Schmitt, J \"o rg Schl \"o tterer, Maurice Van Keulen, and Christin Seifert. 2023. From anecdotal evidence to quantitative evaluation methods: A systematic review on evaluating explainable AI . AC...

    2023

  22. [30]

    W hy should i trust you?

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. https://doi.org/10.1145/2939672.2939778 " W hy should i trust you?": Explaining the predictions of any classifier . In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mini...

    2016

  23. [31]

    Glicksberg, Eyal Zimlichman, Yiftach Barash, Robert Freeman, Alexander W

    Ali Soroush, Benjamin S. Glicksberg, Eyal Zimlichman, Yiftach Barash, Robert Freeman, Alexander W. Charney, Girish N Nadkarni, and Eyal Klang. 2024. https://doi.org/10.1056/AIdbp2300040 Large language models are poor medical coders — benchmarking of medical code querying . NEJ...

    2024 doi

  24. [32]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. https://proceedings.neurips.cc/paper/2017/hash/3f5ee243547dee91fbd053c1c4a845aa-Abstract.html Attention is all you need . In Advances in Neural I...

    2017

  25. [33]

    Sarah Wiegreffe and Ana Marasovi \'c . 2021. https://datasets-benchmarks-proceedings.neurips.cc/paper/2021/hash/698d51a19d8a121ce581499d7b701668-Abstract-round1.html Teach me to explain: A review of datasets for explainable natural language processing . In Proceedings of the N...

    2021

  26. [34]

    Zach Wood-Doughty, Isabel Cachola, and Mark Dredze. 2022. https://doi.org/10.18653/v1/2022.bionlp-1.41 Model distillation for faithful explanations of medical code predictions . In Proceedings of the 21st Workshop on Biomedical Language Processing, pages 412--425, Dublin, Irel...

    2022 doi

  27. [35]

    Zhichao Yang, Sanjit Singh Batra, Joel Stremmel, and Eran Halperin. 2023. https://arxiv.org/abs/2311.13735 Surpassing GPT-4 medical coding with a two-stage approach . Preprint, arXiv:2311.13735

    2023 arXiv

  28. [36]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  29. [37]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.