Pith. sign in

REVIEW 4 major objections 6 minor 81 references

BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read Adding language-segment tokens and medical word embeddings to transformer encoders improves emergency classification of Korean-English code-switched pediatric EMR notes.

desk verdict Sensible framework for Korean-English code-switched EMR triage, but test-set hyperparameter selection and a possible label leak undercut the reported gains; worth a revision, not a reject. read the letter →

arxiv 2412.11671 v1 pith:U7AVX5Y7 submitted 2024-12-16 cs.CL cs.AI

classification cs.CLcs.AI
keywords code-switchingelectronicmedicalrecordspediatricemergencydepartmenttriageclassificationBioSent2VecsegmenttokensmultilingualBERTKorean-EnglishclinicalNLP
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

BioBridge is a fine-tuning framework for transformer encoders that classifies emergency versus non-emergency visits from the free-text Present Illness notes of Korean-English code-switched pediatric EMRs. It adds two mechanisms: language segment tokens that mark whether each span is Korean or English, and a unified bio-embedding that injects English medical word vectors, from a biomedical sentence-embedding model, into the encoder's token representations. In the paper's reported results, BioBridge improves F1, AUROC, and AUPRC over the plain encoder for every backbone tested, and it lowers the Brier score for most backbones; BioBridge-XLM, for instance, gains 0.85% F1, 0.75% AUROC, 0.76% AUPRC, and a 3.04% lower Brier score relative to XLM. The claim matters because code-switched clinical text is common in non-English-speaking hospitals and is poorly served by general-domain multilingual models.

What carries the argument

The load-bearing object is the pair of segment tokens [B-K] and [B-E], which are prepended to the Korean and English spans of each tokenized input sentence so the encoder can treat the two languages as distinct modalities, mirroring how multimodal transformers separate video, text, and audio inputs. The second mechanism is the unified bio-embedding: English subword tokens are re-joined into words, passed through the fixed BioSent2Vec medical feature extractor, and projected by a fully connected layer into the encoder's hidden dimension, where they are combined with the token embeddings. The classification readout is the usual [CLS] embedding, and the framework is trained end-to-end with only the encoder and projection layer updated.

What would settle it

Mask or remove from each test-set Present Illness note every phrase that names a label-defining intervention (blood test, urinalysis, IV, nebulizer, drug administration, admission) and retrain or re-evaluate BioBridge; if its F1, AUROC, and AUPRC advantages over the plain encoder disappear or shrink to near zero, the reported gains were carried by direct documentation of the label rather than by clinical reasoning from the note.

Watch

Extended reading notes

Core claim

The paper's central claim is that code-switched clinical notes can be handled more effectively by treating each language as a separate modality inside a pre-trained transformer and by supplementing the encoder with medical-domain embeddings at the word level. Concretely, BioBridge inserts [B-K] and [B-E] segment tokens so the encoder can tell Korean spans from English spans, and it reconstructs the English subword tokens into whole words, feeds those words through BioSent2Vec to get medical features, and projects those features into the encoder's hidden space. Tested on emergency/non-emergency classification of 87,759 pediatric EMR Present Illness notes, the framework improves all four reported metrics over the corresponding baseline for both Korean-specific encoders (KR-BERT, KoBERT) and multilingual encoders (XLM, mBERT, XLM-R). The authors also show by ablation that each module contributes to the gain.

Load-bearing premise

The whole result depends on the emergency label meaning what the framework is asked to predict: cases are labeled emergency when the record shows the patient received blood tests, urinalysis, IV hydration, nebulizer treatment, immediate drugs, or admission, so if the free-text notes already mention those interventions, the model can learn to spot documented care instead of anticipating clinical need.

Editorial extensions

If this is right

  • If the framework works as reported, code-switched clinical text can be classified with off-the-shelf multilingual encoders plus a lightweight preprocessing step, without biomedical resources like UMLS that are unavailable for Korean.
  • The Brier-score reductions on most backbones suggest better-calibrated probability estimates, which matters for triage support where decision thresholds are set by clinicians.
  • Because the modules attach to any transformer encoder, the same recipe could be reused for other code-switched language pairs; the paper identifies this as its planned next step.
  • The method's gains are additive: ablations show both the segment tokens and the bio-embedding contribute, so hospitals could adopt either module alone if data or compute is limited.

Reading between the lines

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

  • Because the emergency label is defined by whether the patient received interventions such as blood tests, IV hydration, or admission, and the model reads the notes that document those visits, the reported gains may partly reflect detecting documentation of treatment rather than predicting the need for it; a strong test would be to mask intervention-related phrases and see whether the gap survives.
  • The bio-embedding module only covers English medical words, so Korean medical terminology is still handled by the encoder's general-domain Korean knowledge; a Korean medical embedding or a cross-lingual medical feature extractor could plausibly give larger gains.
  • The framework's transferability likely depends on how cleanly the tokenizer splits the non-English language; languages without clear subword boundaries may need different segment markers or a different word-reconstruction rule.
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 / 6 minor

Summary. The paper proposes BioBridge, a two-module framework for classifying emergency versus non-emergency cases from Korean-English code-switched Present Illness (PI) notes in a pediatric emergency department. The first module, 'bridging modality in context,' inserts segment tokens [B-K] and [B-E] to separate Korean and English spans in the input. The second module, 'unified bio-embedding,' extracts English medical word features with a fixed BioSent2Vec encoder, projects them through a fully connected layer, and integrates them into a pre-trained transformer encoder. Experiments on a private dataset of 87,759 PI notes compare four machine-learning baselines and five encoder-based models (KR-BERT, KoBERT, mBERT-cased/uncased, XLM, XLM-R), reporting small F1, AUROC, AUPRC, and Brier-score improvements for BioBridge variants; the headline result is BioBridge-XLM with +0.85% F1 over XLM. An ablation study decomposes the two modules. The paper also describes a preprocessing method that decodes common medical abbreviations and symbols.

Significance. If the findings were valid, BioBridge would be a useful early exploration of segment-token-based language-modality separation and BioSent2Vec fusion for code-switched EMR classification, an understudied problem. The framework is simple, and the paper targets a practically relevant decision-support task with a large private dataset. However, the reported improvements are small (at most 1.1% F1), and the evaluation protocol is compromised as written by explicit test-set hyperparameter selection; no confidence intervals or significance tests are provided. The method definition also omits the exact fusion operation and the handling of multiple code-switches. The paper states that source code will be made public, which is commendable, but the current evidence does not support the central claim of significant improvement.

major comments (4)
  1. [Section IV-F, Tables 2 and 4; Section V] Section IV-F states that a grid search over learning rates was run 'for all experiments' and that 'the optimal hyperparameters for the test set are shown in Table 2'; Section V repeats the same wording for Table 4. Taken literally, this means the test set was used to select hyperparameters, so the results in Tables 3 and 5 are not independent held-out evaluations. The reported gains are small (e.g., BioBridge-XLM +0.85% F1, +0.75% AUROC), and no confidence intervals, error bars, or multiple-seed results are given. Please re-run the experiments selecting hyperparameters on the development set only (or nested cross-validation), report test metrics for the dev-selected configuration, and include at least five seeds and a significance test (e.g., bootstrap or paired test) for the headline comparisons. If the wording is a mistake, correct it and explain the actual selection protocol.
  2. [Section IV-C and Section IV-B] The emergency label is defined by whether the patient received interventions such as blood tests, urinalysis, IV hydration, nebulizer treatment, immediate drugs, or admission (Section IV-C), while the predictive input is the free-text PI note (Section IV-B). The paper does not establish that the PI note was written before these interventions or that intervention-related mentions are excluded from the input. If the notes contain phrases such as 'IV started' or 'admitted', the model can learn to detect documented care rather than predict the need for it, which would invalidate the decision-support claim. Please provide evidence of temporal ordering or a redaction protocol, and report how performance changes when intervention-related terms are removed from the input (e.g., a vocabulary-based ablation).
  3. [Section III-B2, Eq. (7)] The fusion operation for the unified bio-embedding is not specified. Equation (7) only states that f_theta is in R^{m x h_M}, and the text says the projected BioSent2Vec features are 'integrated into' the encoder, but it does not state whether they are added to the token embeddings of English subword tokens, concatenated with the encoder's hidden states, or combined in another way. The position-alignment procedure from m word-level features to the token-level sequence is also missing. Please give the exact fusion equation, including the role of the [B-E] spans, and the alignment rule; without this, the method is not reproducible.
  4. [Section III-B, Eq. (5) and Figure 1] The formal definition of the bridging-modality input, Eq. (5), covers only a single Korean-to-English switch: x_bri = {[CLS], [B-K], [tokens]_kor, [B-E], [tokens]_eng, [SEP]}. The real PI notes shown in Figure 1 (and implied by the statistics in Table 1) contain multiple alternating Korean and English spans, e.g., '환아가 vomiting 10 회로 emergency department 내원'. The paper does not define how [B-K] and [B-E] tokens are inserted for arbitrary multi-switch sentences, nor how the language-span boundaries are detected. Please specify the span-detection and token-insertion algorithm, or restrict the method and experiments to the single-switch case.
minor comments (6)
  1. [Abstract] The sentence 'the proposed BioBridge significantly performance traditional machine learning and pre-trained encoder-based models' contains a grammatical error ('performance' should be 'outperforms'); also, 'significantly' is not supported by significance tests.
  2. [Section IV-E] The F1 threshold is described as set at 0.595 'to align with the label frequency ratio of the label 1 ratio'. Please state explicitly whether this ratio is computed from the training, development, or test set; if it is computed from the test set, it is another test-set-dependent choice affecting the F1 comparison.
  3. [Section V.A] The heading 'Uninifed bio-embedding' contains a typo ('Uninifed' should be 'Unified'). There are also typographical errors elsewhere, including 'MIMMIC-III' in Section II-A (should be 'MIMIC-III') and 'Herhert M Adler' in reference [76]; please proofread the manuscript.
  4. [Tables 2 and 4] The column header 'Param' appears to be an artifact; the tables show only Model, Batch Size, and Learning Rate. Please remove or fill that column so the tables are unambiguous.
  5. [Section IV-G and Table 5] The narrative around Table 5 selectively highlights improvements: for BioBridge-XLM-Rbase the Brier score worsens from 18.78 to 19.59 relative to the XLM-Rbase baseline, and the 'w/ Bio-embedding' variant's Brier score is 23.33. The text says the module 'consistently enhanced performance on almost all metrics' but does not discuss these negative cases; please report and discuss all results transparently.
  6. [Section VII] The conclusion claims 'state-of-the-art performance' but the comparisons are only against internal baselines; no prior code-switched EMR classification systems are compared. Please temper the claim or provide external comparisons.

Circularity Check

2 steps flagged · score 6.0 of 10

The reported test-set results are not independent predictions: learning rates are grid-searched on the test set itself, so Tables 3 and 5 are post-selection numbers.

  1. fitted input called prediction [Section IV-F (Training details on encoder based model); repeated in Section V (Ablation Study)]
    "Also, we used 5 epochs and gridsearch of learning rate ∈ {2e−6, 3e−6, 5e−6, 1e−5, 2e−5, 3e−5, 4e−5, 5e−5, 6e−5} for all experiments. The optimal hyperparameters for the test set are shown in Table 2."

    Table 3 then reports F1, AUROC, AUPRC, and Brier scores on the same test set, and Section V repeats “The optimal hyperparameters for the test set are shown in Table 4” before Table 5. Because the learning rate was selected by test-set performance, the reported test metrics are the result of a search over that same test set, not an evaluation of a fixed model on untouched data. The central claim of consistent improvement (e.g., +0.85% F1, +0.75% AUROC, +0.76% AUPRC for BioBridge-XLM) is therefore a post-selection observation, with no independent dev-only selection, error bars, or significance tests to separate selection effect from method effect.

  2. fitted input called prediction [Section V, Ablation Study]
    "The optimal hyperparameters for the test set are shown in Table 4."

    The ablation comparisons in Table 5 are also based on hyperparameters chosen on the test set, so the incremental contributions attributed to “bridging modality in context” and “unified bio-embedding” are not independent held-out estimates; they are selected on the same labels used for evaluation.

full rationale

Apart from the test-set hyperparameter selection, the BioBridge construction is not circular. The segment-token module is an external multimodal idea (ref. [54]) and the BioSent2Vec feature extractor is an external pre-trained model; the paper does not derive its gains from a self-citation or from an equation that defines the output in terms of the input. The emergency label is defined by interventions, but the paper does not state that PI notes contain those interventions, so the target-leakage possibility is an external validity concern rather than a demonstrated circular step. The dominant issue is evaluation circularity: the phrase “optimal hyperparameters for the test set” appears in both the main training section and the ablation section, and the same test set is then used to produce the headline numbers. That makes the reported predictions partly fitted to the evaluation data. Because the claimed improvements are small and no uncertainty quantification is provided, the central empirical claim is not supported as an independent prediction as written.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No new scientific entities are postulated; the segment tokens [B-K]/[B-E] are model-internal markers, not independent claims about the world. The ledger instead captures the hand-chosen threshold, test-set-tuned learning rates, and the domain assumptions about BioSent2Vec transfer and label validity.

free parameters (2)
  • learning rate per model selected on test set = 2e-6 to 6e-5 depending on model (Table 2)
    Grid-searched learning rates were chosen based on test set performance ('The optimal hyperparameters for the test set are shown in Table 2'), which constitutes fitting model selection to the evaluation data.
  • F1 decision threshold = 0.595
    The F1 threshold is set to 0.595 to align with the label frequency ratio; it is a hand-chosen operating point, though not tuned to optimize test F1.
assumptions (3)
  • domain assumption BioSent2Vec provides useful medical features for English clinical terms in Korean-English code-switched notes
    The unified bio-embedding module relies on BioSent2Vec, pretrained on English biomedical literature and MIMIC-III, to transfer medical knowledge into the encoder. No evaluation of this transfer on Korean-English notes is provided beyond the main results.
  • domain assumption The operational definition of emergency (receipt of blood tests, IV hydration, nebulizer, drugs, or admission) is a valid proxy for emergency status
    Section IV-C defines labels by interventions performed during the PED visit; this assumes those interventions indicate true emergency and that notes do not already encode them, otherwise the model may learn to detect documented care rather than need.
  • domain assumption Code-switched English tokens in these EMRs are the clinically salient modality
    The method focuses medical feature extraction on English words only (Eq. 3), assuming Korean portions carry less clinical signal; this is stated in the intro but not independently validated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR." pith.science (2026). https://pith.science/paper/U7AVX5Y7

@misc{pith2026241211671,
  author       = {Pith},
  title        = {Pith review of: BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U7AVX5Y7}},
  note         = {Machine review of arXiv:2412.11671}
}
read the original abstract

Pediatric Emergency Department (PED) overcrowding presents a significant global challenge, prompting the need for efficient solutions. This paper introduces the BioBridge framework, a novel approach that applies Natural Language Processing (NLP) to Electronic Medical Records (EMRs) in written free-text form to enhance decision-making in PED. In non-English speaking countries, such as South Korea, EMR data is often written in a Code-Switching (CS) format that mixes the native language with English, with most code-switched English words having clinical significance. The BioBridge framework consists of two core modules: "bridging modality in context" and "unified bio-embedding." The "bridging modality in context" module improves the contextual understanding of bilingual and code-switched EMRs. In the "unified bio-embedding" module, the knowledge of the model trained in the medical domain is injected into the encoder-based model to bridge the gap between the medical and general domains. Experimental results demonstrate that the proposed BioBridge significantly performance traditional machine learning and pre-trained encoder-based models on several metrics, including F1 score, area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), and Brier score. Specifically, BioBridge-XLM achieved enhancements of 0.85% in F1 score, 0.75% in AUROC, and 0.76% in AUPRC, along with a notable 3.04% decrease in the Brier score, demonstrating marked improvements in accuracy, reliability, and prediction calibration over the baseline XLM model. The source code will be made publicly available.

Figures

Figures reproduced from arXiv: 2412.11671 by the authors.

Figure 1
Figure 1. FIGURE 1 [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. FIGURE 2 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

81 extracted references · 64 canonical work pages

  1. [1]

    Emergency department crowding and hospital bed shortage: is lean a smart answer? a systematic review

    Sabina Bucci, AG De Belvis, S Marventano, AC De Leva, Maria Tanzariello, Maria Lucia Specchia, Walter Ricciardi, and Francesco Franceschi. Emergency department crowding and hospital bed shortage: is lean a smart answer? a systematic review. Eur Rev Med Pharmacol Sci , 20(20):4209–19, 2016

  2. [2]

    Time series model for forecasting the number of new admission inpatients

    Lingling Zhou, Ping Zhao, Dongdong Wu, Cheng Cheng, and Hao Huang. Time series model for forecasting the number of new admission inpatients. BMC medical informatics and decision making , 18:1–11, 2018

  3. [3]

    Outcomes of crowding in emergency departments; a systematic review

    Hamid Reza Rasouli, Ali Aliakbar Esfahani, Mohammad Nobakht, Mohsen Eskandari, Sardollah Mahmoodi, Hassan Goodarzi, and Mohsen Abbasi Farajzadeh. Outcomes of crowding in emergency departments; a systematic review. Archives of academic emergency medicine, 7(1), 2019

  4. [4]

    Overcrowding in emergency department: causes, consequences, and solutions—a narrative review

    Marina Sartini, Alessio Carbone, Alice Demartini, Luana Giribone, Mar- tino Oliva, Anna Maria Spagnolo, Paolo Cremonesi, Francesco Canale, and Maria Luisa Cristina. Overcrowding in emergency department: causes, consequences, and solutions—a narrative review. In Healthcare, vol- ume 10, page 1625. MDPI, 2022

  5. [5]

    Overcrowding crisis in our nation’s emergency departments: is our safety net unraveling? Pediatrics, 114(3):878–888, 2004

    American Academy of Pediatrics Committee on Pediatric Emer- gency Medicine et al. Overcrowding crisis in our nation’s emergency departments: is our safety net unraveling? Pediatrics, 114(3):878–888, 2004

  6. [6]

    Understanding the characteristics of recurrent visits to the emergency department by paediatric patients: a retrospective observa- VOLUME 11, 2023 9 J

    Byung-Soo Kim, Jung-Y oun Kim, Sung-Hyuk Choi, and Y oung-Hoon Y oon. Understanding the characteristics of recurrent visits to the emergency department by paediatric patients: a retrospective observa- VOLUME 11, 2023 9 J. Jeon et al.: BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR. tional study conducted at three tertiary hos...

  7. [7]

    The role of the electronic medical record (emr) in care delivery development in developing countries: a systematic review

    Faustine Williams and Suzanne Austin Boren. The role of the electronic medical record (emr) in care delivery development in developing countries: a systematic review. Informatics in primary care , 16(2), 2008

  8. [8]

    Word2vec inversion and traditional text classifiers for phenotyping lupus

    Clayton A Turner, Alexander D Jacobs, Cassios K Marques, James C Oates, Diane L Kamen, Paul E Anderson, and Jihad S Obeid. Word2vec inversion and traditional text classifiers for phenotyping lupus. BMC medical informatics and decision making , 17(1):1–11, 2017

Show all 81 references
  1. [9]

    Mak- ing a case in medical work: implications for the electronic medical record

    Mark Hartswood, Rob Procter, Mark Rouncefield, and Roger Slack. Mak- ing a case in medical work: implications for the electronic medical record. Computer Supported Cooperative Work (CSCW), 12:241–266, 2003

  2. [10]

    Benefits and drawbacks of elec- tronic health record systems

    Nir Menachemi and Taleah H Collum. Benefits and drawbacks of elec- tronic health record systems. Risk management and healthcare policy , pages 47–55, 2011

  3. [11]

    Prediction of emergency department patient disposition based on natural language processing of triage notes

    Nicholas W Sterling, Rachel E Patzer, Mengyu Di, and Justin D Schrager. Prediction of emergency department patient disposition based on natural language processing of triage notes. International journal of medical informatics, 129:184–188, 2019

  4. [12]

    A large language model for electronic health records

    Xi Y ang, Aokun Chen, Nima PourNejatian, Hoo Chang Shin, Kaleb E Smith, Christopher Parisien, Colin Compas, Cheryl Martin, Anthony B Costa, Mona G Flores, et al. A large language model for electronic health records. NPJ digital medicine, 5(1):194, 2022

  5. [13]

    Use of natural language processing of patient-initiated electronic health record messages to identify patients with covid-19 infection

    Kellen Mermin-Bunnell, Y uanda Zhu, Andrew Hornback, Gregory Damhorst, Tiffany Walker, et al. Use of natural language processing of patient-initiated electronic health record messages to identify patients with covid-19 infection. jama network open 6, 7 (07 2023), e2322299– e23...

  6. [14]

    Predicting of anaphylaxis in big data emr by exploring machine learning approaches

    Isabel Segura-Bedmar, Cristobal Colon-Ruiz, Miguél Ángel Tejedor- Alonso, and Mar Moro-Moro. Predicting of anaphylaxis in big data emr by exploring machine learning approaches. Journal of biomedical informatics, 87:50–59, 2018

  7. [15]

    Au- tomatic classification of online doctor reviews: evaluation of text classifier algorithms

    Ryan Rivas, Niloofar Montazeri, Nhat XT Le, and V agelis Hristidis. Au- tomatic classification of online doctor reviews: evaluation of text classifier algorithms. Journal of medical Internet research, 20(11):e11141, 2018

  8. [16]

    An automated system for identifying alcohol use status from clinical text

    Hadeel Alzoubi, Naeem Ramzan, Raid Alzubi, and Ehsan Mesbahi. An automated system for identifying alcohol use status from clinical text. In 2018 International Conference on Computing, Electronics & Communica- tions Engineering (iCCECE) , pages 41–46. IEEE, 2018

  9. [17]

    Effi- cient estimation of word representations in vector space

    Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. Effi- cient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781, 2013

  10. [18]

    Multi-layer representation learning for medical concepts

    Edward Choi, Mohammad Taha Bahadori, Elizabeth Searles, Catherine Coffey, Michael Thompson, James Bost, Javier Tejedor-Sojo, and Jimeng Sun. Multi-layer representation learning for medical concepts. In proceed- ings of the 22nd ACM SIGKDD international conference on knowledge ...

  11. [19]

    Biowordvec, improving biomedical word embeddings with subword infor- mation and mesh

    Yijia Zhang, Qingyu Chen, Zhihao Y ang, Hongfei Lin, and Zhiyong Lu. Biowordvec, improving biomedical word embeddings with subword infor- mation and mesh. Scientific data, 6(1):52, 2019

  12. [20]

    Biobert based named entity recognition in electronic medical record

    Xin Y u, Wenshen Hu, Sha Lu, Xiaoyan Sun, and Zhenming Y uan. Biobert based named entity recognition in electronic medical record. In 2019 10th international conference on information technology in medicine and education (ITME), pages 49–52. IEEE, 2019

  13. [21]

    Med-bert: pretrained contextualized embeddings on large-scale structured electronic health records for disease prediction

    Laila Rasmy, Y ang Xiang, Ziqian Xie, Cui Tao, and Degui Zhi. Med-bert: pretrained contextualized embeddings on large-scale structured electronic health records for disease prediction. NPJ digital medicine, 4(1):86, 2021

  14. [22]

    Publicly available clinical bert embeddings

    Emily Alsentzer, John R Murphy, Willie Boag, Wei-Hung Weng, Di Jin, Tristan Naumann, and Matthew McDermott. Publicly available clinical bert embeddings. arXiv preprint arXiv:1904.03323, 2019

  15. [23]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018

  16. [24]

    Overcoming the challenges of unstructured data in multisite, electronic medical record- based abstraction

    Brock Polnaszek, Andrea Gilmore-Bykovskyi, Melissa Hovanes, Rachel Roiland, Patrick Ferguson, Roger Brown, and Amy JH Kind. Overcoming the challenges of unstructured data in multisite, electronic medical record- based abstraction. Medical care, 54(10):e65–e72, 2016

  17. [25]

    Challenges in understanding clinical notes: Why nlp engines fall short and where background knowledge can help

    Sujan Perera, Amit Sheth, Krishnaprasad Thirunarayan, Suhas Nair, and Neil Shah. Challenges in understanding clinical notes: Why nlp engines fall short and where background knowledge can help. In Proceedings of the 2013 international workshop on Data management & analytics for...

  18. [26]

    Role and challenges of unstructured big data in healthcare

    Kiran Adnan, Rehan Akbar, Siak Wang Khor, and Adnan Bin Amanat Ali. Role and challenges of unstructured big data in healthcare. Data Management, Analytics and Innovation: Proceedings of ICDMAI 2019, V olume 1, pages 301–323, 2020

  19. [27]

    Machine learning in medicine: a practical introduction to natural language processing

    Conrad J Harrison and Chris J Sidey-Gibbons. Machine learning in medicine: a practical introduction to natural language processing. BMC medical research methodology, 21(1):1–11, 2021

  20. [28]

    High-throughput phenotyping with electronic medical record data using a common semi-supervised approach (phecap)

    Yichi Zhang, Tianrun Cai, Sheng Y u, Kelly Cho, Chuan Hong, Jiehuan Sun, Jie Huang, Y uk-Lam Ho, Ashwin N Ananthakrishnan, Zongqi Xia, et al. High-throughput phenotyping with electronic medical record data using a common semi-supervised approach (phecap). Nature protocols, 14(...

  21. [29]

    V alidation of electronic health record phenotyp- ing of bipolar disorder cases and controls

    Victor M Castro, Jessica Minnier, Shawn N Murphy, Isaac Kohane, Su- sanne E Churchill, Vivian Gainer, Tianxi Cai, Alison G Hoffnagle, Y ael Dai, Stefanie Block, et al. V alidation of electronic health record phenotyp- ing of bipolar disorder cases and controls. American Journa...

  22. [30]

    Current status of electronic medical record systems in hospitals and clinics in korea

    Y oung-Taek Park and Dongwoon Han. Current status of electronic medical record systems in hospitals and clinics in korea. Healthcare informatics research, 23(3):189–198, 2017

  23. [31]

    Electronic medical records and machine learning in approaches to drug development

    Ayaka Shinozaki. Electronic medical records and machine learning in approaches to drug development. In Artificial intelligence in Oncology drug discovery and development . IntechOpen, 2020

  24. [32]

    Consistency across the hierarchies of the umls semantic network and metathesaurus

    James J Cimino, Hua Min, and Y ehoshua Perl. Consistency across the hierarchies of the umls semantic network and metathesaurus. Journal of biomedical informatics, 36(6):450–461, 2003

  25. [33]

    The unified medical language system (umls): integrat- ing biomedical terminology

    Olivier Bodenreider. The unified medical language system (umls): integrat- ing biomedical terminology. Nucleic acids research , 32(suppl_1):D267– D270, 2004

  26. [34]

    The uni- fied medical language system

    Donald AB Lindberg, Betsy L Humphreys, and Alexa T McCray. The uni- fied medical language system. Yearbook of medical informatics, 2(01):41– 51, 1993

  27. [35]

    The neighborhood auditing tool: a hybrid interface for auditing the umls

    C Paul Morrey, James Geller, Michael Halper, and Y ehoshua Perl. The neighborhood auditing tool: a hybrid interface for auditing the umls. Jour- nal of biomedical informatics , 42(3):468–489, 2009

  28. [36]

    Xgboost: A scalable tree boosting system

    Tianqi Chen and Carlos Guestrin. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining , pages 785–794, 2016

  29. [37]

    The regression analysis of binary sequences

    David R Cox. The regression analysis of binary sequences. Journal of the Royal Statistical Society Series B: Statistical Methodology, 20(2):215–232, 1958

  30. [38]

    Greedy function approximation: a gradient boosting machine

    Jerome H Friedman. Greedy function approximation: a gradient boosting machine. Annals of statistics, pages 1189–1232, 2001

  31. [39]

    Random forests

    Leo Breiman. Random forests. Machine learning, 45:5–32, 2001

  32. [40]

    SKTBrain. Kobert. https://github.com/SKTBrain/KoBERT

  33. [41]

    Kr-bert: A small-scale korean-specific language model

    Sangah Lee, Hansol Jang, Y unmee Baik, Suzi Park, and Hyopil Shin. Kr-bert: A small-scale korean-specific language model. arXiv preprint arXiv:2008.03979, 2020

  34. [42]

    Cross-lingual language model pretraining

    Alexis Conneau and Guillaume Lample. Cross-lingual language model pretraining. Advances in neural information processing systems, 32, 2019

  35. [43]

    Unsupervised cross-lingual represen- tation learning at scale

    Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and V eselin Stoyanov. Unsupervised cross-lingual represen- tation learning at scale. In Dan Jurafsky, Joyce Chai, Natalie Schlut...

  36. [44]

    Korean clinical entity recognition from diagnosis text using bert

    Y oung-Min Kim and Tae-Hoon Lee. Korean clinical entity recognition from diagnosis text using bert. BMC Medical Informatics and Decision Making, 20:1–9, 2020

  37. [45]

    Keyword extrac- tion algorithm for classifying smoking status from unstructured bilingual electronic health records based on natural language processing

    Y e Seul Bae, Kyung Hwan Kim, Han Kyul Kim, Sae Won Choi, Taehoon Ko, Hee Hwa Seo, Hae-Y oung Lee, and Hyojin Jeon. Keyword extrac- tion algorithm for classifying smoking status from unstructured bilingual electronic health records based on natural language processing. Applied...

  38. [46]

    Code-switching in conversation: Language, interaction and identity

    Peter Auer. Code-switching in conversation: Language, interaction and identity. Routledge, 2013

  39. [47]

    F oundations of bilingual education and bilingualism

    Colin Baker. F oundations of bilingual education and bilingualism. Multi- lingual matters, 2011

  40. [48]

    Language and advertis- ing effectiveness: Code-switching in the korean marketplace

    Jungsun Ahn, Carrie La Ferle, and Doohwang Lee. Language and advertis- ing effectiveness: Code-switching in the korean marketplace. International Journal of Advertising, 36(3):477–495, 2017

  41. [49]

    Korean/English intrasentential code-switching: Matrix language assignment and linguistic constraints

    Jun-Eon Park. Korean/English intrasentential code-switching: Matrix language assignment and linguistic constraints . University of Illinois at Urbana-Champaign, 1990. 10 VOLUME 11, 2023 J. Jeon et al.: BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR

  42. [50]

    Intrasentential codeswitching in japanese and en- glish

    NISHIMURA Miwa. Intrasentential codeswitching in japanese and en- glish. Ph. D dissertation. University of Pennsylvania , 1985

  43. [51]

    Addressing code-switching in french/algerian arabic speech

    Djegdjiga Amazouz, Martine Adda-Decker, and Lori Lamel. Addressing code-switching in french/algerian arabic speech. In Interspeech 2017 , pages 62–66, 2017

  44. [52]

    Chinese clinical named entity recognition with variant neural structures based on bert methods

    Xiangyang Li, Huan Zhang, and Xiao-Hua Zhou. Chinese clinical named entity recognition with variant neural structures based on bert methods. Journal of biomedical informatics , 107:103422, 2020

  45. [53]

    Honglei Liu, Zhiqiang Zhang, Y an Xu, Ni Wang, Y anqun Huang, Zhenghan Y ang, Rui Jiang, and Hui Chen. Use of bert (bidirectional encoder repre- sentations from transformers)-based deep learning method for extracting evidences in chinese radiology reports: development of a com...

  46. [54]

    Bridging text and video: A universal multimodal transformer for audio-visual scene- aware dialog

    Zekang Li, Zongjia Li, Jinchao Zhang, Y ang Feng, and Jie Zhou. Bridging text and video: A universal multimodal transformer for audio-visual scene- aware dialog. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 29:2476–2483, 2021

  47. [55]

    Scibert: A pretrained language model for scientific text

    Iz Beltagy, Kyle Lo, and Arman Cohan. Scibert: A pretrained language model for scientific text. arXiv preprint arXiv:1903.10676, 2019

  48. [56]

    Aligning books and movies: Towards story-like visual explanations by watching movies and reading books

    Y ukun Zhu, Ryan Kiros, Rich Zemel, Ruslan Salakhutdinov, Raquel Ur- tasun, Antonio Torralba, and Sanja Fidler. Aligning books and movies: Towards story-like visual explanations by watching movies and reading books. In Proceedings of the IEEE international conference on comput...

  49. [57]

    Ccnet: Extracting high quality monolingual datasets from web crawl data

    Guillaume Wenzek, Marie-Anne Lachaux, Alexis Conneau, Vishrav Chaudhary, Francisco Guzmán, Armand Joulin, and Edouard Grave. Ccnet: Extracting high quality monolingual datasets from web crawl data. arXiv preprint arXiv:1911.00359, 2019

  50. [58]

    A pre-trained bert for korean medical natural language processing

    Y oojoong Kim, Jong-Ho Kim, Jeong Moon Lee, Moon Joung Jang, Y un Jin Y um, Seongtae Kim, Unsub Shin, Y oung-Min Kim, Hyung Joon Joo, and Sanghoun Song. A pre-trained bert for korean medical natural language processing. Scientific Reports, 12(1):13847, 2022

  51. [59]

    Definitions and assessment approaches for emergency medical services for children

    Susan Fuchs, Mark Terry, Kathleen Adelgais, Marlene Bokholdt, Jane Brice, Kathleen M Brown, Arthur Cooper, Mary E Fallat, Katherine E Remick, Keith Widmeier, et al. Definitions and assessment approaches for emergency medical services for children. Pediatrics, 138(6), 2016

  52. [60]

    Predicting hospital readmission for lupus patients: An rnn-lstm-based deep-learning methodology

    Bhargava K Reddy and Dursun Delen. Predicting hospital readmission for lupus patients: An rnn-lstm-based deep-learning methodology. Computers in biology and medicine , 101:199–209, 2018

  53. [61]

    A novel deep learning–based system for triage in the emergency department using electronic medical records: Retrospective cohort study

    Li-Hung Y ao, Ka-Chun Leung, Chu-Lin Tsai, Chien-Hua Huang, and Li- Chen Fu. A novel deep learning–based system for triage in the emergency department using electronic medical records: Retrospective cohort study. Journal of Medical Internet Research , 23(12):e27008, 2021

  54. [62]

    Unsupervised learn- ing of sentence embeddings using compositional n-gram features

    Matteo Pagliardini, Prakhar Gupta, and Martin Jaggi. Unsupervised learn- ing of sentence embeddings using compositional n-gram features. arXiv preprint arXiv:1703.02507, 2017

  55. [63]

    Biosentvec: creating sentence embeddings for biomedical texts

    Qingyu Chen, Yifan Peng, and Zhiyong Lu. Biosentvec: creating sentence embeddings for biomedical texts. In 2019 IEEE International Conference on Healthcare Informatics (ICHI) , pages 1–5. IEEE, 2019

  56. [64]

    Named entity recogni- tion method of chinese emr based on bert-bilstm-crf

    Wenchao Gao, Xiaohui Zheng, and Shanshan Zhao. Named entity recogni- tion method of chinese emr based on bert-bilstm-crf. In Journal of physics: conference series, volume 1848, page 012083. IOP Publishing, 2021

  57. [65]

    Unsupervised cross-lingual represen- tation learning at scale

    Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and V eselin Stoyanov. Unsupervised cross-lingual represen- tation learning at scale. arXiv preprint arXiv:1911.02116, 2019

  58. [66]

    Identifying alcohol-related information from unstructured bilingual clinical notes with multilingual transformers

    Han Kyul Kim, Y ujin Park, Y eju Park, Eunji Choi, Sodam Kim, Hahyun Y ou, and Y e Seul Bae. Identifying alcohol-related information from unstructured bilingual clinical notes with multilingual transformers. IEEE Access, 11:16066–16075, 2023

  59. [67]

    Deep learning-based natural language processing for detecting medical symptoms and histories in emergency patient triage

    Siryeol Lee, Juncheol Lee, Juntae Park, Jiwoo Park, Dohoon Kim, Joohyun Lee, and Jaehoon Oh. Deep learning-based natural language processing for detecting medical symptoms and histories in emergency patient triage. The American Journal of Emergency Medicine , 77:29–38, 2024

  60. [68]

    De-identification of clinical notes with pseudo- labeling using regular expression rules and pre-trained bert

    Jiyong An, Jiyun Kim, Leonard Sunwoo, Hyunyoung Baek, Sooyoung Y oo, and Seunggeun Lee. De-identification of clinical notes with pseudo- labeling using regular expression rules and pre-trained bert. 2023

  61. [69]

    Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks

    Jiasen Lu, Dhruv Batra, Devi Parikh, and Stefan Lee. Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks. Advances in neural information processing systems , 32, 2019

  62. [70]

    Lxmert: Learning cross-modality encoder rep- resentations from transformers

    Hao Tan and Mohit Bansal. Lxmert: Learning cross-modality encoder rep- resentations from transformers. arXiv preprint arXiv:1908.07490, 2019

  63. [71]

    Videobert: A joint model for video and language representation learning

    Chen Sun, Austin Myers, Carl V ondrick, Kevin Murphy, and Cordelia Schmid. Videobert: A joint model for video and language representation learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 7464–7473, 2019

  64. [72]

    Unified vision-language pre-training for image captioning and vqa

    Luowei Zhou, Hamid Palangi, Lei Zhang, Houdong Hu, Jason Corso, and Jianfeng Gao. Unified vision-language pre-training for image captioning and vqa. In Proceedings of the AAAI conference on artificial intelligence , volume 34, pages 13041–13049, 2020

  65. [73]

    Dynamic fusion with intra-and inter- modality attention flow for visual question answering

    Peng Gao, Zhengkai Jiang, Haoxuan Y ou, Pan Lu, Steven CH Hoi, Xi- aogang Wang, and Hongsheng Li. Dynamic fusion with intra-and inter- modality attention flow for visual question answering. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , ...

  66. [74]

    Progressive attention memory network for movie story question answering

    Junyeong Kim, Minuk Ma, Kyungsu Kim, Sungjin Kim, and Chang D Y oo. Progressive attention memory network for movie story question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8337–8346, 2019

  67. [75]

    Modality shifting attention network for multi-modal video question answering

    Junyeong Kim, Minuk Ma, Trung Pham, Kyungsu Kim, and Chang D Y oo. Modality shifting attention network for multi-modal video question answering. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10106–10115, 2020

  68. [76]

    The history of the present illness as treatment: who’s listening, and why does it matter? The Journal of the American Board of Family Practice, 10(1):28–35, 1997

    Herhert M Adler. The history of the present illness as treatment: who’s listening, and why does it matter? The Journal of the American Board of Family Practice, 10(1):28–35, 1997

  69. [77]

    Roberta: A robustly optimized bert pretraining approach

    Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and V eselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692, 2019

  70. [78]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems , 32, 2019

  71. [79]

    Huggingface’s transformers: State-of-the-art natural language processing

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Rémi Louf, Morgan Funtowicz, et al. Huggingface’s transformers: State-of-the-art natural language processing. arXiv preprint arXiv:1910.03771, 2019

  72. [80]

    A unified approach to interpreting model predictions

    Scott M Lundberg and Su-In Lee. A unified approach to interpreting model predictions. Advances in neural information processing systems, 30, 2017. VOLUME 11, 2023 11

  73. [2020]

    Association for Computational Linguistics

Pith tools

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