Pith. sign in

REVIEW 4 major objections 5 minor 37 references

Cohort Retrieval using Dense Passage Retrieval

T0 review · 4 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 A custom-trained dense passage retriever, built from echocardiography reports, outperforms BM25 and off-the-shelf clinical embeddings on patient cohort retrieval tasks.

desk verdict A useful applied DPR study for echo cohort retrieval, but evaluation overlaps training queries and the numbers overstate generalization. read the letter →

arxiv 2507.01049 v1 pith:QX7G5C2J submitted 2025-06-26 cs.IR cs.CL

Pranav Jadhav This is my paper
classification cs.IRcs.CL
keywords densepassageretrievalpatientcohortechocardiographyelectronichealthrecordsmultiplenegativesrankinglossBM25LVEFclinicalinformation
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 patient cohort retrieval from echocardiography reports is better solved by a dense passage retriever trained from scratch on the domain's own query-passage pairs than by traditional lexical retrieval or off-the-shelf clinical embedding models. It builds a dataset from a 43,472-report intensive-care echocardiography collection by turning Left Ventricle findings into query statements and report summaries into passages, then evaluates on held-out, paraphrased, out-of-distribution, and numerical LVEF queries. On the non-OOD tasks the custom model reaches P@10 between 0.82 and 0.99, clearly above BM25 and ClinicalBERT. If these results hold, they establish a template for adapting DPR to other medical note domains with dense, condition-rich text.

What carries the argument

The central object is a modified Multiple Negatives Ranking Loss applied to a shared bi-encoder DPR model. Standard MNR treats every other pair's positive passage as an in-batch negative, which is wrong when many queries share relevant passages; the paper sets the cosine similarity of non-anchor positives to minus infinity so they contribute no misleading gradient. Training triplets draw the positive and a hard negative from different subcategories of the same clinical condition, forcing the model to learn fine-grained distinctions, and numerical LVEF training is supported by a temporary database of extracted values, signs, and ranges.

What would settle it

Construct a fresh set of LVEF queries written by clinicians in natural language, with no overlap with the paper's templates, and compare the v1 and v2 models against BM25 on the same passage corpus. If P@10 drops to BM25's level, the reported numerical superiority is an artifact of template memorization. The same test can be run on paraphrased queries by generating them with a different LLM and prompt than the one used to create the training paraphrases.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that a custom-trained DPR embedding model is superior to traditional and off-the-shelf state-of-the-art methods for patient cohort retrieval in echocardiography. The authors convert unstructured echo reports into a query-passage corpus, then train a shared bi-encoder embedding model from scratch so that queries and report summaries are close in embedding space when the summary mentions the queried condition. On evaluation sets built to mimic clinical use, the custom model beats both BM25 and off-the-shelf clinical and general embedding models on held-out reports, paraphrased queries, and LVEF numerical queries: the numerical variant reaches P@10 of 0.99, the held-out variant reaches 0.85, and paraphrased-query performance reaches 0.68. The paper also reports that out-of-distribution rare conditions remain unsolved, with all dense variants scoring near zero while BM25 reaches 0.48.

Load-bearing premise

The load-bearing premise is that the evaluation sets measure true clinical retrieval behavior rather than the model recognizing query templates it already saw in training.

Editorial extensions

If this is right

  • Domain-specific DPR can replace term matching and general clinical embeddings for cohort retrieval when the same clinical condition is expressed in many surface forms.
  • The modified Multiple Negatives Ranking Loss, which masks other pairs' positives, makes DPR training viable when one query maps to many relevant passages.
  • Numerical cohort queries such as LVEF thresholds can be handled by combining heuristic value extraction with templated training queries, reaching near-perfect P@10 on the paper's evaluation.
  • The same dataset-construction and evaluation recipe transfers to other medical note domains that have structured findings plus free-text summaries.
  • Out-of-distribution rare conditions remain an open problem: lexical BM25 stays the best method there, and dense variants score at most 0.06 P@10.

Reading between the lines

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

  • The numerical evaluation uses queries generated from the same templates seen in training, so the 0.98-0.99 P@10 likely overstates performance on free-text numerical questions; varying the wording or adding distractors would test this.
  • The held-out set reuses training query statements with new passages, so part of the gain may come from memorized query-to-condition mappings rather than passage understanding; the paraphrased set is the cleaner evidence of generalization.
  • The OOD collapse suggests a hybrid system that combines lexical retrieval for rare terms with dense retrieval for common paraphrases may be the practical path, though the paper does not propose it.
  • If the approach is extended beyond the Left Ventricle to other anatomies, the condition hierarchy and hard-negative sampling would need to handle overlapping labels, possibly requiring a different loss.
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

4 major / 5 minor

Summary. The paper applies Dense Passage Retrieval (DPR) to patient cohort retrieval in echocardiography. The authors transform the Left Ventricle section of MIMIC-III echocardiography reports into query statements and use the Summary sections as passages, constructing an inverted index for relevance labels. They evaluate three custom-trained DPR variants (v0, v1, v2), BM25, ClinicalBERT, and Jina embeddings on four tasks: held-out passages, paraphrased queries, numerical LVEF retrieval, and out-of-distribution (OOD) conditions. They report that the custom models outperform baselines on the first two tasks, achieve near-perfect scores on the numerical task for v1/v2, and all fail on OOD, while BM25 performs best on OOD. They also propose a modification to Multiple Negatives Ranking loss that masks in-batch positives that are relevant to the anchor query.

Significance. If the reported results are valid, the paper makes a useful practical contribution: a systematic pipeline for converting an unstructured clinical EHR corpus into a query-passage retrieval dataset, a domain-specific dense retriever that outperforms off-the-shelf alternatives on several retrieval tasks, and a loss modification that addresses a subtle problem in cohort retrieval with multiple relevant passages. The paper is honest in reporting the OOD failure and in stating the limitations of the dataset. The weaknesses are concentrated in the evaluation design: the held-out set reuses training queries, the numerical evaluation queries come from the same template distribution used in training, and there are no statistical significance or uncertainty estimates. These issues weaken the external validity of the headline claim of 'superior performance' but do not invalidate the relative ordering of the models on the specific test splits.

major comments (4)
  1. [§3.3.1] The held-out set is described as containing 'the same set of queries used in the training set' with only new passages. This design tests generalization to new passages for query types already seen in training, but not generalization to novel query phrasings. Consequently, the high P@10 values in Table 3 (0.85 for v0) could largely reflect the model memorizing query-to-condition mappings rather than matching new query formulations. The four subcategories excluded from training are the only genuinely unseen query types, yet their results are not reported separately. Please report performance on those four subcategories separately, or construct a held-out set with queries not seen during training, to support the claim of superior generalization to clinical queries.
  2. [§3.3.3 and §4] The numerical evaluation queries are automatically generated from templated expressions shown in Table 2 (e.g., 'LVEF {sign} {mid_value}%') with random values, and the v1 training set is augmented with triplets created from the same temporary LVEF database and similar templates. Although the paper states that the evaluation queries were excluded from training, the distribution of the evaluation queries is essentially identical to the training distribution. The near-perfect P@10 scores of 0.98 and 0.99 for v1 and v2 are therefore expected for a model trained on that exact distribution. To support the claim of superior numerical retrieval, evaluate on free-text numerical queries written by clinicians or on templates that were not used in training; otherwise the numerical results should be described as in-distribution performance, not generalization.
  3. [§3.2 and §3.4] Relevance labels are constructed by exact statement occurrence via the inverted index: a passage is relevant only if it contains the exact statement (or one of its manually grouped variations) assigned to the query. A passage that expresses the same clinical condition using different wording is counted as negative. This labeling scheme penalizes models that generalize semantically and may advantage models that rely on surface-form matching. As a result, the absolute P@10 and R-Precision values in Table 3 are not directly interpretable as clinical retrieval accuracy. The paper should either add a small set of manual/clinical relevance judgments to calibrate the automatic labels, or explicitly discuss this limitation and its potential effect on the reported scores.
  4. [Table 3] All reported metrics are single-point estimates with no error bars, bootstrap confidence intervals, or significance tests. Given the small evaluation sets (100 numerical queries, 60 paraphrased queries, and a held-out set whose size is not stated), differences such as v0 vs. v1 on the held-out set (P@10 0.85 vs. 0.81) may be within noise. The paper should report variance across multiple training runs or bootstrap confidence intervals, and ideally paired significance tests for the main comparisons, to support the claim that the custom DPR model is 'superior' to the baselines.
minor comments (5)
  1. [§4] The paper says the model was 'trained from scratch without any prior pretrained weights' but also that it used 'msmarco-bert-base-dot-v5 as architecture.' Since 'msmarco-bert-base-dot-v5' refers to a pretrained checkpoint, clarify whether the model was initialized with random weights (i.e., only the architecture is taken from that model) or with the pretrained checkpoint; the current wording is contradictory.
  2. [§3.3.2] The OOD evaluation uses conditions that occur only in single-digit numbers, and the trained models achieve P@10 of 0.00–0.06. Please report the number of OOD queries and passages used in the evaluation, as results at this scale are highly variable and a single query can shift the metric substantially.
  3. [Table 3] The model names in Table 3 are concatenated with the first metric values (e.g., 'echo-retriever-model v00.85'); add spacing for readability.
  4. [Limitations] The Limitations section states that 'comparisons were restricted to embedding-based models,' but BM25 is a lexical, non-embedding baseline. Please revise the wording to acknowledge that the comparison includes a traditional IR baseline.
  5. [§3.3.4] The paraphrased query set contains 60 queries. Given the small size, report the number of queries per subcategory and indicate whether the paraphrases were manually validated for medical correctness and preservation of intent.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported retrieval scores are empirical evaluations on held-out passages, not derived predictions; the evaluation design is standard and does not reduce to training inputs by construction.

full rationale

The paper's central claim is that a custom-trained DPR model outperforms baselines on echocardiography cohort retrieval. This claim is supported by measured P@10, P@100, and R-Precision values on evaluation sets described in Section 3.3. There is no equation-level derivation in which an output is defined in terms of the target quantity or in which a fitted parameter is renamed as a prediction. The held-out set (Section 3.3.1) uses the same query strings as training but an entirely different set of passages; this is a standard passage-retrieval evaluation design that tests generalization to new documents, not a circular reduction. The numerical task (Section 3.3.3) generates evaluation queries with random values and signs, so the specific query instances are unseen even though the templates resemble training templates; this tests value-level generalization rather than memorization of fixed query strings. The OOD evaluation (Section 3.3.2) explicitly removes rare conditions from training, and the paper honestly reports poor OOD performance, which further indicates that the evaluation is not rigged to force the claimed result. The loss modification in Equation 1 is a standard adaptation of Multiple Negatives Ranking Loss, not a self-referential construction. There are no load-bearing self-citations, no imported uniqueness theorems, and no ansatz smuggled in via citation. The concern that high scores partly reflect query-to-condition memorization is a generalizability/external-validity limitation, not circularity: the model is trained to map query statements to passages, and evaluating on new passages containing the same conditions is the intended task. The paper does not claim to predict unseen query types, and where it tests unseen query types (OOD), it reports weak performance. Thus no circular step meeting the required evidentiary standard is present.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central empirical results rest on the curated query-passage dataset and the chosen hyperparameters. The most significant unverified inputs are the clinical accuracy of the heuristics-generated annotations and the assumption that summary passages alone capture all relevant cohort criteria. No new theoretical entities are introduced.

free parameters (5)
  • learning_rate = 2e-5
    Used for DPR fine-tuning; chosen from sentence-transformers default, not fit to test metric.
  • batch_size = 64
    Training batch size; affects in-batch negative sampling and was chosen by the authors.
  • max_seq_length = 300
    Set based on token length distribution in Table 1; truncates longer passages and queries.
  • warmup_steps = 1000
    Learning rate warmup; standard value, selected from the sentence-transformers training script.
  • training_steps = 200000
    Number of training steps for DPR fine-tuning; chosen by the authors without reported convergence analysis.
assumptions (3)
  • domain assumption Echocardiography report summary sections contain sufficient and consistent information to serve as retrieval passages for all query conditions.
    Passages are the Summary sections of 42,996 reports (Section 3.2). If a condition appears in Findings but is absent from Summary, the query-passage relevance mapping would be incomplete, yet this is not validated.
  • domain assumption The automated heuristic plus manual supervision annotation correctly maps statements to clinical conditions and subcategories.
    All query-passage relevance judgments derive from these annotations, which are not released (Section 3.1). The paper does not report inter-annotator agreement or validation against clinician gold labels.
  • domain assumption The MNR loss modification (masking other in-batch positives) does not introduce degenerate learning dynamics beyond the intended effect.
    The authors do not ablate the loss modification or analyze its impact on gradient behavior; they state it avoids misleading gradients but provide no empirical verification of this claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cohort Retrieval using Dense Passage Retrieval." pith.science (2026). https://pith.science/paper/QX7G5C2J

@misc{pith2026250701049,
  author       = {Pith},
  title        = {Pith review of: Cohort Retrieval using Dense Passage Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QX7G5C2J}},
  note         = {Machine review of arXiv:2507.01049}
}
read the original abstract

Patient cohort retrieval is a pivotal task in medical research and clinical practice, enabling the identification of specific patient groups from extensive electronic health records (EHRs). In this work, we address the challenge of cohort retrieval in the echocardiography domain by applying Dense Passage Retrieval (DPR), a prominent methodology in semantic search. We propose a systematic approach to transform an echocardiographic EHR dataset of unstructured nature into a Query-Passage dataset, framing the problem as a Cohort Retrieval task. Additionally, we design and implement evaluation metrics inspired by real-world clinical scenarios to rigorously test the models across diverse retrieval tasks. Furthermore, we present a custom-trained DPR embedding model that demonstrates superior performance compared to traditional and off-the-shelf SOTA methods.To our knowledge, this is the first work to apply DPR for patient cohort retrieval in the echocardiography domain, establishing a framework that can be adapted to other medical domains.

Figures

Figures reproduced from arXiv: 2507.01049 by the authors.

Figure 1
Figure 1. Echocardiographic report with queries high [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Systematic mapping of conditions, subcategories, statements, and associated passage IDs [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

37 extracted references · 22 canonical work pages

  1. [1]

    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 INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Satya Almasian, Milena Bruseva, and Michael Gertz. 2024. https://doi.org/10.18653/v1/2024.findings-emnlp.707 Numbers matter! bringing quantity-awareness to retrieval systems . In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 12120--12136, Miami, Florida, USA. Association for Computational Linguistics

    doi:10.18653/v1/2024.findings-emnlp.707 2024

  4. [4]

    Cimino, John Ely, and Hong Yu

    YongGang Cao, Feifan Liu, Pippa Simpson, Lamont Antieau, Andrew Bennett, James J. Cimino, John Ely, and Hong Yu. 2011. https://doi.org/10.1016/j.jbi.2011.01.004 AskHERMES : An online question answering system for complex clinical questions . Journal of Biomedical Informatics, 44(2):277--288

    work page doi:10.1016/j.jbi.2011.01.004 2011

  5. [5]

    Cohen, Yanshan Wang, Andrew Wen, Sijia Liu, Hongfang Liu, and William R

    Steven Chamberlin, Steven Bedrick, Arjeh M. Cohen, Yanshan Wang, Andrew Wen, Sijia Liu, Hongfang Liu, and William R. Hersh. 2019. https://api.semanticscholar.org/CorpusID:208101747 Evaluation of patient-level retrieval from electronic health record data for a cohort discovery task . JAMIA Open, 3:395 -- 404

    work page 2019

  6. [6]

    Eun Suk Chang and Javed Mostafa. 2022. https://api.semanticscholar.org/CorpusID:256665900 Cohort identification from free-text clinical notes using snomed ct's hierarchical semantic relations . AMIA ... Annual Symposium proceedings. AMIA Symposium, 2022:349--358

    work page 2022

  7. [7]

    is this document relevant?…probably

    Fabio Crestani, Mounia Lalmas, Cornelis J. Van Rijsbergen, and Iain Campbell. 1998. https://doi.org/10.1145/299917.299920 “is this document relevant?…probably”: a survey of probabilistic models in information retrieval . ACM Comput. Surv., 30(4):528–552

    work page arXiv 1998

  8. [8]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. https://arxiv.org/abs/1810.04805 Bert: Pre-training of deep bidirectional transformers for language understanding . Preprint, arXiv:1810.04805

    arXiv 2019

Show all 37 references
  1. [9]

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, et al. 2024. https://arxiv.org/abs/2407.21783 The llama 3 herd of models . arXiv preprint arXiv:2407.21783

    2024 arXiv

  2. [10]

    Jean-Baptiste Excoffier, Tom Roehr, Alexei Figueroa, Jens-Michalis Papaioannou, Keno Kyrill Bressem, and Matthieu Ortala. 2024. https://api.semanticscholar.org/CorpusID:266755594 Generalist embedding models are better at short-context clinical semantic search than specialized ...

    2024 arXiv

  3. [11]

    Johnson, Khader Shameer, Li Li, Rongbing Chen, and Joel T

    Benjamin Scott Glicksberg, Riccardo Miotto, Kipp W. Johnson, Khader Shameer, Li Li, Rongbing Chen, and Joel T. Dudley. 2018. https://api.semanticscholar.org/CorpusID:21430063 Automated disease cohort selection using word embeddings from electronic health records . Pacific Symp...

    2018

  4. [12]

    Goodwin and Sanda M

    Travis R. Goodwin and Sanda M. Harabagiu. 2018. https://api.semanticscholar.org/CorpusID:53740285 Learning relevance models for patient cohort retrieval . JAMIA Open, 1:265 -- 275

    2018

  5. [13]

    Lucas, Naoto Usuyama, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, and Hoifung Poon

    Yu Gu, Robert Tinn, Hao Cheng, Michael R. Lucas, Naoto Usuyama, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, and Hoifung Poon. 2020. https://api.semanticscholar.org/CorpusID:220919723 Domain-specific language model pretraining for biomedical natural language processing . ACM T...

    2020

  6. [14]

    Michael G \"u nther, Jackmin Ong, Isabelle Mohr, Alaeddine Abdessalem, Tanguy Abel, Mohammad Kalim Akram, Susana Guzman, Georgios Mastrapas, Saba Sturua, Bo Wang, Maximilian Werk, Nan Wang, and Han Xiao. 2024. https://arxiv.org/abs/2310.19923 Jina embeddings 2: 8192-token gene...

    2024 arXiv

  7. [15]

    Shashank Gupta. 2023. https://api.semanticscholar.org/CorpusID:260704560 Top k relevant passage retrieval for biomedical question answering . ArXiv, abs/2308.04028

    2023 arXiv

  8. [16]

    Matthew Henderson, Rami Al-Rfou, Brian Strope, Yun hsuan Sung, Laszlo Lukacs, Ruiqi Guo, Sanjiv Kumar, Balint Miklos, and Ray Kurzweil. 2017. https://arxiv.org/abs/1705.00652 Efficient natural language response suggestion for smart reply . Preprint, arXiv:1705.00652

    2017 arXiv

  9. [17]

    Kexin Huang, Jaan Altosaar, and Rajesh Ranganath. 2020. https://arxiv.org/abs/1904.05342 Clinicalbert: Modeling clinical notes and predicting hospital readmission . Preprint, arXiv:1904.05342

    2020 arXiv

  10. [18]

    Tian Kang, Shaodian Zhang, Youlan Tang, Gregory William Hruby, Alex Rusanov, No \'e mie Elhadad, and Chunhua Weng. 2017. https://api.semanticscholar.org/CorpusID:3740574 Eliie: An open-source information extraction system for clinical trial eligibility criteria . Journal of th...

    2017

  11. [19]

    Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.550 Dense passage retrieval for open-domain question answering . In Proceedings of the 2020 Conference on Empiric...

    2020 doi

  12. [20]

    Antoine Kossaify and Gilles Grollier. 2014. https://api.semanticscholar.org/CorpusID:9720356 Echocardiography practice: Insights into appropriate clinical use, technical competence and quality improvement program . Clinical Medicine Insights: Cardiology, 8:1 -- 7

    2014

  13. [21]

    Mendoza, Petr Knoth, Gabriella Pasi, and Allan Hanbury

    Wojciech Kusa, \'O scar E. Mendoza, Petr Knoth, Gabriella Pasi, and Allan Hanbury. 2023. https://api.semanticscholar.org/CorpusID:262091275 Effective matching of patients to clinical trials using entity extraction and neural re-ranking . Journal of biomedical informatics, page 104444

    2023

  14. [22]

    Gloria Hyunjung Kwak, Dana Moukheiber, Mira Moukheiber, Lama Moukheiber, Sulaiman Moukheiber, Neel Butala, Leo Anthony Celi, and Christina Chen. 2024. https://doi.org/10.13026/xhrz-ht59 Echonotes structured database derived from mimic-iii (echo-note2num)

    2024 doi

  15. [23]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2021. https://arxiv.org/abs/2005.11401 Retrieval-augmented generation for knowledge-int...

    2021 arXiv

  16. [24]

    Ning Liu, Qian Hu, Huayun Xu, Xing Xu, and Mengxin Chen. 2022. https://doi.org/10.1109/TII.2021.3131180 Med-bert: A pretraining framework for medical records named entity recognition . IEEE Transactions on Industrial Informatics, 18(8):5600--5608

    2022

  17. [25]

    Hersh, and Hongfang Liu

    Sijia Liu, Yanshan Wang, Andrew Wen, Liwei Wang, Na Hong, Feichen Shen, Steven Bedrick, William R. Hersh, and Hongfang Liu. 2019. https://api.semanticscholar.org/CorpusID:59158901 Create: Cohort retrieval enhanced by analysis of text from electronic health records using omop c...

    2019 arXiv

  18. [26]

    Finelli, Greg L

    Xiong Liu, Luca A. Finelli, Greg L. Hersch, and Iya G. Khalil. 2020. https://api.semanticscholar.org/CorpusID:229331857 Attention-based lstm network for covid-19 clinical trial parsing . 2020 IEEE International Conference on Big Data (Big Data), pages 3761--3766

    2020

  19. [27]

    Man Luo, Arindam Mitra, Tejas Gokhale, and Chitta Baral. 2022. https://api.semanticscholar.org/CorpusID:246035178 Improving biomedical information retrieval with neural retrievers . In AAAI Conference on Artificial Intelligence

    2022

  20. [28]

    Masoud Monajatipoor, Jiaxin Yang, Joel Stremmel, Melika Emami, Fazlolah Mohaghegh, Mozhdeh Rouhsedaghat, and Kai-Wei Chang. 2024. https://arxiv.org/abs/2404.07376 Llms in biomedicine: A study on clinical named entity recognition . Preprint, arXiv:2404.07376

    2024 arXiv

  21. [29]

    Nye, Jay DeYoung, Eric P

    Benjamin E. Nye, Jay DeYoung, Eric P. Lehman, Ani Nenkova, Iain James Marshall, and Byron C. Wallace. 2020. https://api.semanticscholar.org/CorpusID:222177200 Understanding clinical trial reports: Extracting medical entities and their relations . AMIA ... Annual Symposium proc...

    2020

  22. [30]

    Nils Reimers and Iryna Gurevych. 2019. https://arxiv.org/abs/1908.10084 Sentence-bert: Sentence embeddings using siamese bert-networks . Preprint, arXiv:1908.10084

    2019 arXiv

  23. [31]

    Gonz \'a lez, and Manuel Montes-y G \'o mez

    Andr \'e s Rosso-Mateus, Fabio A. Gonz \'a lez, and Manuel Montes-y G \'o mez. 2020. https://doi.org/10.18653/v1/2020.coling-main.548 A deep metric learning method for biomedical passage retrieval . In Proceedings of the 28th International Conference on Computational Linguisti...

    2020 doi

  24. [32]

    Shivade, Preethi Raghavan, Eric Fosler-Lussier, Peter J

    Chaitanya P. Shivade, Preethi Raghavan, Eric Fosler-Lussier, Peter J. Emb \'i , No \'e mie Elhadad, Stephen B. Johnson, and Albert M. Lai. 2013. https://api.semanticscholar.org/CorpusID:1841217 A review of approaches to identifying patient phenotype cohorts using electronic he...

    2013

  25. [33]

    Sarvesh Soni and Kirk Roberts. 2020. https://api.semanticscholar.org/CorpusID:221640646 Patient cohort retrieval using transformer language models . AMIA ... Annual Symposium proceedings. AMIA Symposium, 2020:1150--1159

    2020

  26. [34]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2023. https://arxiv.org/abs/1706.03762 Attention is all you need . Preprint, arXiv:1706.03762

    2023 arXiv

  27. [35]

    Wagholikar, Christina M

    Kavishwar B. Wagholikar, Christina M. Fischer, Alyssa P. Goodson, Christopher Herrick, Martin Rees, Eloy Toscano, Calum A. Macrae, Benjamin Morgan Scirica, Akshay S. Desai, and Shawn N. Murphy. 2018. https://api.semanticscholar.org/CorpusID:52822404 Extraction of ejection frac...

    2018

  28. [36]

    Stephen Wu, Kirk Roberts, Surabhi Datta, Jingcheng Du, Zongcheng Ji, Yuqi Si, Sarvesh Soni, Qiong Wang, Qiang Wei, Yang Xiang, Bo Zhao, and Hua Xu. 2020. https://doi.org/10.1093/jamia/ocz200 Deep learning in clinical natural language processing: a methodical review . Journal o...

    2020 doi

  29. [37]

    Glass, and Jimeng Sun

    Xingyao Zhang, Cao Xiao, Lucas M. Glass, and Jimeng Sun. 2020. https://arxiv.org/abs/2001.08179 Deepenroll: Patient-trial matching with deep embedding and entailment prediction . Preprint, arXiv:2001.08179

    2020 arXiv

Pith tools

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