Pith. sign in

REVIEW 4 major objections 4 minor 43 references

CLINICSUM: Utilizing Language Models for Generating Clinical Summaries from Patient-Doctor Conversations

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

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

Pith's one-line read The paper claims that pairing retrieval-based filtering with a fine-tuned open language model generates SOAP clinical summaries that beat GPT-4-class models on automatic metrics and clinician preference.

desk verdict Useful dataset, sensible pipeline, but the superiority claim rests on a confounded evaluation. read the letter →

arxiv 2412.04254 v1 pith:YJDZEKNA submitted 2024-12-05 cs.CL cs.AI

Subash Neupane , Himanshu Tripathi , Shaswata Mitra , Sean Bozorgzad , Sudip Mittal , Shahram Rahimi , Amin Amirlatifi This is my paper · ORCID
classification cs.CLcs.AI
keywords clinicalsummarizationSOAPnotespatient-doctorconversationslargelanguagemodelsretrieval-basedfilteringparameter-efficientfine-tuningLoRAzero-shot
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 tries to establish that a two-stage pipeline can turn raw patient-doctor conversation transcripts into structured SOAP clinical summaries more reliably than calling a general-purpose chat model directly. The first stage retrieves only the sentences that carry subjective, objective, assessment, and plan content, and the second stage, a language model fine-tuned on 1,473 expert-validated examples, writes the summary. If correct, this matters because it offers a path to automated clinical documentation that is both more accurate and more affordable than relying on proprietary GPT-class models.

What carries the argument

The load-bearing mechanism is the separation of extraction from generation. A retriever-based filter splits the transcript into sentences, embeds and indexes them, and uses an ensemble of a sparse retriever (BM25) and a dense retriever with Reciprocal Rank Fusion to pull out only chunks containing subjective, objective, assessment, and plan information. The compressed context then feeds an inference module: a 4-bit quantized, LoRA-fine-tuned causal language model prompted in a zero-shot setting, which writes the SOAP summary from the filtered chunks plus an instruction template.

What would settle it

Run CLINICSUM alongside GPT-4-class baselines on real, de-identified patient-doctor transcripts with independently written clinician SOAP notes, then have clinicians blind-compare the summaries for factual errors and missing critical information; the central claim collapses if CLINICSUM no longer wins. A simpler check is to recompute ROUGE and BERTScore using reference summaries written by clinicians from scratch instead of model-generated, SME-corrected summaries.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central finding is that CLINICSUM, retrieval-based filtering followed by zero-shot inference from a fine-tuned open-source language model, outperforms GPT-4-Turbo, GPT-4-O-Mini, and GPT-3.5-Turbo on ROUGE-1, ROUGE-2, ROUGE-L, and BERTScore, and that four clinician subject-matter experts preferred its summaries over the best GPT model's output 39 to 24 with 17 ties. The best configuration pairs the framework with LLaMA-3-8B (ROUGE-1 F1 0.70, ROUGE-L F1 0.55, BERTScore F1 0.84), while the strongest GPT baseline trails on every metric.

Load-bearing premise

The finding rests on 20 simulated conversations played by theater students, and on reference summaries that were drafted by the same kind of model that the framework is compared against; if either the staging or the reference style is not representative of real clinical interactions, the reported advantage may not survive in practice.

Editorial extensions

If this is right

  • Automatic SOAP note generation from conversation transcripts is feasible with open, locally deployed models, avoiding per-call API costs and patient-data sharing with external services.
  • Filtering the transcript before generation reduces token count and, the paper argues, lowers the risk of hallucination and token overflow compared with summarizing full transcripts.
  • Fine-tuning with LoRA on a single 24 GB consumer GPU is enough to reach strong clinical summarization quality, making deployment plausible in resource-constrained healthcare settings.
  • Models whose output token counts stay close to the ground truth tend to score higher, suggesting that summary length control is a useful diagnostic for faithfulness.

Reading between the lines

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

  • Inference: because the reference summaries themselves were produced by a GPT model and then human-corrected, the automatic-metric gap may partly reflect stylistic mimicry of that reference style rather than purely medical quality; a human review focused on factual errors against the original transcript would separate these.
  • Inference: the same retrieval-filtering idea should transfer to other note formats, such as history-and-physical or discharge notes, and to non-English conversations, provided enough validated examples exist for fine-tuning.
  • Inference: a cheap, falsifiable extension is to ablate the retriever, feeding the same fine-tuned model the full transcript in window-sized chunks, to quantify how much of the gain comes from filtering rather than from fine-tuning alone.
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 / 4 minor

Summary. The paper proposes CLINICSUM, a two-module framework for generating SOAP-format clinical summaries from patient-doctor conversations. The first module applies an ensemble retriever (BM25 plus a dense retriever, fused with reciprocal rank fusion) to filter transcript chunks containing Subjective, Objective, Assessment, and Plan content. The second module fine-tunes an open-source PLM with LoRA/QLoRA on 1,473 conversation-summary pairs; these pairs were drafted by GPT-4-O-Mini and then reviewed and corrected by SMEs. Evaluation is carried out on 20 simulated conversations transcribed with Whisper, using ROUGE, BERTScore, and SME preference judgments. The paper reports that CLINICSUM with LLAMA-3-8B outperforms GPT-4-Turbo, GPT-4-O-Mini, and GPT-3.5-Turbo on all automatic metrics and is preferred over GPT-4-O-Mini in the human study.

Significance. If the central claim is sound, the contribution is practically valuable: it combines retrieval-based filtering with parameter-efficient fine-tuning, releases a publicly available training dataset of 1,473 conversation-summary pairs, and demonstrates that a 8B open-weight model can be fine-tuned on a single 24GB consumer GPU. The zero-shot inference setting is a strength, and the inclusion of SME-corrected training summaries is a positive step. However, the evidence as presented does not yet support the paper's strong comparative claims. The evaluation-reference provenance is the load-bearing risk, exactly as the stress-test note identifies: if the 20 test references come from the same GPT-4-O-Mini-plus-SME pipeline used for training, the ROUGE/BERTScore advantages may reflect style mimicry rather than clinical superiority. The small test set, lack of abalation for the retrieval module, absence of significance testing, and a human study that omits a base-model control together mean that the results should be treated as preliminary until the evaluation is strengthened.

major comments (4)
  1. [V-B, V-A] The provenance of the 20 reference summaries used for automatic and human evaluation is not described anywhere in Section V-B. Because Section V-A explicitly says the 1,473 training summaries were drafted by GPT-4-O-Mini and then SME-corrected, if the evaluation references were produced through the same pipeline, the ROUGE and BERTScore comparisons in Table V primarily measure how closely each system reproduces the reference style of that pipeline rather than clinical quality. Please state how the evaluation references were created and, if they share the GPT-4-O-Mini draft stage, either re-score against independently created clinician-written references or report a metric that controls for style overlap (e.g., meaning-based expert annotation).
  2. [V-B, Table V] All automatic results rest on 20 simulated conversations, and Table V reports no confidence intervals or significance tests. With n=20 and four CLINICSUM variants plus three GPT baselines, the observed gaps (e.g., ROUGE-1 F1 0.70 vs 0.64, BERTScore 0.84 vs 0.77) may be within sampling noise; report per-conversation variance, paired tests (e.g., bootstrap or Wilcoxon), and multiplicity correction before claiming general superiority.
  3. [IV, V-B, Table VI] The design does not isolate the contribution of either module. No comparison is made between CLINICSUM-LLAMA-3-8B and the base LLAMA-3-8B (with or without retrieval filtering), and the SME evaluation in Table VI compares only CLINICSUM-LLAMA-3-8B against GPT-4-O-Mini. Consequently, the higher scores cannot be attributed specifically to the retrieval-based filtering or to fine-tuning; an ablation (full system, no retriever, base model zero-shot, retriever plus base model) is needed to support the architecture claims in Sections IV and VII.
  4. [V-B2] The human evaluation, while a useful complement, is not decisive: it uses four SMEs on 20 conversations, excludes ties from win rates (Table VI), reports only moderate IRR (kappa = 0.437, alpha = 0.445 in Table VII), and provides no significance testing. The 61% versus 39% aggregate preference, after excluding 17 ties, should be reported with confidence intervals and inter-rater agreement on each criterion (critical information, factual correctness, irrelevant details) rather than only overall preference.
minor comments (4)
  1. [Throughout] Typos such as 'dataset of consisting' (Abstract), 'figsahre.com' (Section I), 'ground-turth' (Section V-B1), and 'evalaution' (Table VI caption) should be corrected.
  2. [IV-A] Hyperparameters for the retriever ensemble (k, WSparse, WDense, lambda) and the exact embedding model used for dense retrieval are not specified; include them for reproducibility.
  3. [IV-B1] LoRA rank, alpha, dropout, learning rate, batch size, number of epochs, and 4-bit quantization settings are omitted; these are necessary to reproduce Table V.
  4. [V-A] The description of selecting '1,201 clean conversations' from MTS-Dialog needs explicit filtering criteria, and the statement that MTS-Dialog summaries 'average less than three sentences' should be tied to the statistics in Table II.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the automatic-metric comparison is a standard held-out supervised evaluation, with only minor self-citations and an undisclosed test-reference provenance caveat.

full rationale

The claimed derivation is not circular by construction. The PLM is fine-tuned on 1,473 GPT-4-O-Mini-drafted, SME-corrected summaries (Section V-A) and evaluated on 20 newly staged, ASR-transcribed conversations (Section V-B) that are not part of the training set. ROUGE and BERTScore are computed against held-out summaries, and the human preference test independently compares CLINICSUM-LLaMA-3-8B with GPT-4-O-Mini (Table VI), giving CLINICSUM a 61% win rate. The self-citations ([1], [34]) are contextual or used only to name an example dense retriever; they do not carry the argument. The only caveat is that the paper never states how the 20 evaluation ground-truth summaries were produced: Section V-A describes GPT-4-O-Mini-plus-SME generation only for the 1,473 training pairs, while Section V-B.2 refers to 'the same set of 20 conversations and summaries' and to 'ground-truth data' without a provenance statement. If those 20 references came from the same GPT-4-O-Mini-plus-SME pipeline, part of the fine-tuned model's ROUGE/BERTScore advantage could reflect learned mimicry of the reference generator's style rather than clinical superiority. Because the provenance is unstated, this is a validity and completeness limitation, not an exhibited by-construction equivalence; the human evaluation provides independent grounding, so the paper does not reduce to its inputs.

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

The paper's main dependency is on unspecified hyperparameters and on the validity of its self-built dataset and simulated evaluation. It does not introduce new theoretical entities.

free parameters (4)
  • top-k retrieved chunks
    The number of chunks passed to the generator is not specified in the paper; it controls how much context the model sees and thus affects accuracy.
  • retriever weights WSparse, WDense
    Equation (10) assigns weights to sparse and dense retrievers, but the paper does not report the values used.
  • RRF constant lambda
    Equation (11) uses lambda to avoid division by zero; the value is unspecified.
  • LoRA hyperparameters (rank, alpha, dropout)
    The fine-tuning configuration is described only as LoRA on 4-bit quantized weights; specifics needed for replication are missing.
assumptions (4)
  • domain assumption SOAP components can be extracted by retrieving sentences with a generic prompt
    Section IV-A assumes that a sentence-level retriever with prompt 'Extract subjective, objective, assessment, and plan details' identifies all relevant content.
  • domain assumption SME-corrected GPT-4-O-Mini summaries are reliable ground truth
    Section V-A uses these as training targets; Section V-B appears to use similar references for evaluation. If SME corrections were incomplete, the references carry model bias.
  • domain assumption Simulated conversations are representative of real clinical encounters
    Section V-B uses 20 role-played conversations; the authors themselves list this as a limitation in Section VI.
  • standard math Standard retrieval math (BM25, dense embeddings, RRF) is applicable to dialogue chunks
    The framework relies on standard IR techniques without modification; these are well-established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CLINICSUM: Utilizing Language Models for Generating Clinical Summaries from Patient-Doctor Conversations." pith.science (2026). https://pith.science/paper/YJDZEKNA

@misc{pith2026241204254,
  author       = {Pith},
  title        = {Pith review of: CLINICSUM: Utilizing Language Models for Generating Clinical Summaries from Patient-Doctor Conversations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YJDZEKNA}},
  note         = {Machine review of arXiv:2412.04254}
}
read the original abstract

This paper presents ClinicSum, a novel framework designed to automatically generate clinical summaries from patient-doctor conversations. It utilizes a two-module architecture: a retrieval-based filtering module that extracts Subjective, Objective, Assessment, and Plan (SOAP) information from conversation transcripts, and an inference module powered by fine-tuned Pre-trained Language Models (PLMs), which leverage the extracted SOAP data to generate abstracted clinical summaries. To fine-tune the PLM, we created a training dataset of consisting 1,473 conversations-summaries pair by consolidating two publicly available datasets, FigShare and MTS-Dialog, with ground truth summaries validated by Subject Matter Experts (SMEs). ClinicSum's effectiveness is evaluated through both automatic metrics (e.g., ROUGE, BERTScore) and expert human assessments. Results show that ClinicSum outperforms state-of-the-art PLMs, demonstrating superior precision, recall, and F-1 scores in automatic evaluations and receiving high preference from SMEs in human assessment, making it a robust solution for automated clinical summarization.

Figures

Figures reproduced from arXiv: 2412.04254 by the authors.

Figure 1
Figure 1. A graphical overview of the CLINICSUM. P denotes the Patient and D denotes the Doctor in the conversation transcript. S, O, A, and P refer to the Subjective, Objective, Assessment, and Plan components of the clinical summary. provided by healthcare practitioners almost immediately [3] and misconstrue nearly half of what they remember [4]. For healthcare providers, generating these summaries auto￾matically can allevi… view at source ↗
Figure 2
Figure 2. A is graphical illustration of the CLINICSUM architecture. It comprises two modules: retrieved-based filtering and inference. B represents patient-doctor conversation, and C represents generated clinical summary. [...] (used for brevity) indicates that there is more textual information. (C), usually employing traditional keyword-based methods or indexing techniques. We employ BM25 as our sparse retriever (RSparse) t… view at source ↗
Figure 3
Figure 3. as our instruction. On the other hand, prompt directs a fine-tuned PLM to produce clinical summaries in a zero-shot setting. The fine-tuned generator processes the prompt, patient context, and instructions to generate a comprehensive clinical SOAP summary. In the following subsections we describe our fine-tuning approach and then detail summary generation. 1) Fine-tuning: In this work, we leverage PEFT [20] ap￾proac… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Heatmap illustrating the preferences between sum [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Radar chart illustrating how different models compare [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 21 canonical work pages

  1. [1]

    Medinsight: A multi-source context augmentation framework for generating patient-centric medical responses using large language models,

    S. Neupane, S. Mitra, S. Mittal, N. A. Golilarz, S. Rahimi, and A. Amir- latifi, “Medinsight: A multi-source context augmentation framework for generating patient-centric medical responses using large language models,” arXiv preprint arXiv:2403.08607 , 2024

    arXiv 2024

  2. [2]

    Retrieval-augmented generation for ai-generated content: A survey,

    P. Zhao, H. Zhang, Q. Yu, Z. Wang, Y . Geng, F. Fu, L. Yang, W. Zhang, and B. Cui, “Retrieval-augmented generation for ai-generated content: A survey,” arXiv preprint arXiv:2402.19473 , 2024

    arXiv 2024

  3. [3]

    Remembering what the doctor said: organization and adults’ memory for medical information,

    L. C. Mcguire, “Remembering what the doctor said: organization and adults’ memory for medical information,” Experimental aging research, vol. 22, no. 4, pp. 403–428, 1996

    work page 1996

  4. [4]

    Patient information recall in a rheumatology clinic,

    J. L. Anderson, S. Dodman, M. Kopelman, and A. Fleming, “Patient information recall in a rheumatology clinic,” Rheumatology, vol. 18, no. 1, pp. 18–18, 1979

    work page 1979

  5. [5]

    Burnout and doctors: prevalence, prevention and interven- tion,

    S. Kumar, “Burnout and doctors: prevalence, prevention and interven- tion,” in Healthcare, vol. 4, no. 3. MDPI, 2016, p. 37

    work page 2016

  6. [6]

    User-Driven Research of Medical Note Generation Software

    T. Knoll, F. Moramarco, A. P. Korfiatis, R. Young, C. Ruffini, M. Perera, C. Perstl, E. Reiter, A. Belz, and A. Savkov, “User-driven research of medical note generation software,” arXiv preprint arXiv:2205.02549 , 2022

    work page Pith review arXiv 2022

  7. [7]

    A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions,

    L. Huang, W. Yu, W. Ma, W. Zhong, Z. Feng, H. Wang, Q. Chen, W. Peng, X. Feng, B. Qin et al. , “A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions,” arXiv preprint arXiv:2311.05232 , 2023

    arXiv 2023

  8. [8]

    An empirical study of clinical note generation from doctor-patient encounters,

    A. Ben Abacha, W.-w. Yim, Y . Fan, and T. Lin, “An empirical study of clinical note generation from doctor-patient encounters,” in Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics . Dubrovnik, Croatia: Association for Computational Linguistics, May 2023, pp. 2291–2302. [Online]. Available: https:/...

    work page 2023

Show all 43 references
  1. [9]

    Leveraging pretrained models for au- tomatic summarization of doctor-patient conversations,

    L. Zhang, R. Negrinho, A. Ghosh, V . Jagannathan, H. R. Hassanzadeh, T. Schaaf, and M. R. Gormley, “Leveraging pretrained models for au- tomatic summarization of doctor-patient conversations,” arXiv preprint arXiv:2109.12174, 2021

    2021 arXiv

  2. [10]

    Wanglab at mediqa-chat 2023: Clinical note generation from doctor-patient conversations using large language models,

    J. Giorgi, A. Toma, R. Xie, S. S. Chen, K. R. An, G. X. Zheng, and B. Wang, “Wanglab at mediqa-chat 2023: Clinical note generation from doctor-patient conversations using large language models,” arXiv preprint arXiv:2305.02220, 2023

    2023 arXiv

  3. [11]

    Language mod- els are few-shot learners,

    T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell et al., “Language mod- els are few-shot learners,” Advances in neural information processing systems, vol. 33, pp. 1877–1901, 2020

    1901

  4. [12]

    Towards an automated soap note: classi- fying utterances from medical conversations,

    B. Schloss and S. Konam, “Towards an automated soap note: classi- fying utterances from medical conversations,” in Machine Learning for Healthcare Conference. PMLR, 2020, pp. 610–631

    2020

  5. [13]

    Allocation of physician time in ambulatory practice: a time and motion study in 4 specialties,

    C. Sinsky, L. Colligan, L. Li, M. Prgomet, S. Reynolds, L. Goeders, J. Westbrook, M. Tutty, and G. Blike, “Allocation of physician time in ambulatory practice: a time and motion study in 4 specialties,” Annals of internal medicine , vol. 165, no. 11, pp. 753–760, 2016

    2016

  6. [14]

    The problem oriented record as a basic tool in medical education, patient care and clinical research

    L. Weed, “The problem oriented record as a basic tool in medical education, patient care and clinical research.”Annals of clinical research, vol. 3, no. 3, pp. 131–134, 1971

    1971

  7. [15]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems , vol. 30, 2017

    2017

  8. [16]

    Recent advances in natural language processing via large pre-trained language models: A survey,

    B. Min, H. Ross, E. Sulem, A. P. B. Veyseh, T. H. Nguyen, O. Sainz, E. Agirre, I. Heintz, and D. Roth, “Recent advances in natural language processing via large pre-trained language models: A survey,” ACM Computing Surveys, vol. 56, no. 2, pp. 1–40, 2023

    2023

  9. [17]

    Survey on factuality in large language models: Knowledge, retrieval and domain-specificity,

    C. Wang, X. Liu, Y . Yue, X. Tang, T. Zhang, C. Jiayang, Y . Yao, W. Gao, X. Hu, Z. Qi et al. , “Survey on factuality in large language models: Knowledge, retrieval and domain-specificity,” arXiv preprint arXiv:2310.07521, 2023

    arXiv

  10. [18]

    Overcoming catastrophic forgetting in neural networks,

    J. Kirkpatrick, R. Pascanu, N. Rabinowitz, J. Veness, G. Desjardins, A. A. Rusu, K. Milan, J. Quan, T. Ramalho, A. Grabska-Barwinska et al. , “Overcoming catastrophic forgetting in neural networks,” Pro- ceedings of the national academy of sciences , vol. 114, no. 13, pp. 3521...

    2017

  11. [19]

    Retrieval- augmented generation for knowledge-intensive nlp tasks,

    P. Lewis, E. Perez, A. Piktus, F. Petroni, V . Karpukhin, N. Goyal, H. Küttler, M. Lewis, W.-t. Yih, T. Rocktäschel et al. , “Retrieval- augmented generation for knowledge-intensive nlp tasks,” Advances in Neural Information Processing Systems , vol. 33, pp. 9459–9474, 2020

    2020

  12. [20]

    Parameter-efficient transfer learning for nlp,

    N. Houlsby, A. Giurgiu, S. Jastrzebski, B. Morrone, Q. De Laroussilhe, A. Gesmundo, M. Attariyan, and S. Gelly, “Parameter-efficient transfer learning for nlp,” in International conference on machine learning . PMLR, 2019, pp. 2790–2799

    2019

  13. [21]

    The power of scale for parameter-efficient prompt tuning,

    B. Lester, R. Al-Rfou, and N. Constant, “The power of scale for parameter-efficient prompt tuning,” arXiv preprint arXiv:2104.08691 , 2021

    2021 arXiv

  14. [22]

    Dylora: Parameter efficient tuning of pre-trained models using dynamic search- free low-rank adaptation,

    M. Valipour, M. Rezagholizadeh, I. Kobyzev, and A. Ghodsi, “Dylora: Parameter efficient tuning of pre-trained models using dynamic search- free low-rank adaptation,” arXiv preprint arXiv:2210.07558 , 2022

    2022 arXiv

  15. [23]

    Qlora: Efficient finetuning of quantized llms,

    T. Dettmers, A. Pagnoni, A. Holtzman, and L. Zettlemoyer, “Qlora: Efficient finetuning of quantized llms,” Advances in Neural Information Processing Systems, vol. 36, 2024

    2024

  16. [24]

    Abstractive dialogue summarization with sentence-gated modeling optimized by dialogue acts,

    C.-W. Goo and Y .-N. Chen, “Abstractive dialogue summarization with sentence-gated modeling optimized by dialogue acts,” in 2018 IEEE Spoken Language Technology Workshop (SLT) . IEEE, 2018, pp. 735– 742

    2018

  17. [25]

    Keep meeting summaries on topic: Abstractive multi-modal meeting summarization,

    M. Li, L. Zhang, H. Ji, and R. J. Radke, “Keep meeting summaries on topic: Abstractive multi-modal meeting summarization,” in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019, pp. 2190–2196

    2019

  18. [26]

    Learning to summarize radiology findings,

    Y . Zhang, D. Y . Ding, T. Qian, C. D. Manning, and C. P. Lan- glotz, “Learning to summarize radiology findings,” arXiv preprint arXiv:1809.04698, 2018

    2018 arXiv

  19. [27]

    An automated medical scribe for documenting clinical encounters,

    G. Finley, E. Edwards, A. Robinson, M. Brenndoerfer, N. Sadoughi, J. Fone, N. Axtmann, M. Miller, and D. Suendermann-Oeft, “An automated medical scribe for documenting clinical encounters,” in Proceedings of the 2018 Conference of the North American Chapter of the Association ...

    2018

  20. [28]

    Generating medical reports from patient-doctor conversations using sequence-to-sequence models,

    S. Enarvi, M. Amoia, M. D.-A. Teba, B. Delaney, F. Diehl, S. Hahn, K. Harris, L. McGrath, Y . Pan, J. Pinto et al. , “Generating medical reports from patient-doctor conversations using sequence-to-sequence models,” in Proceedings of the first workshop on natural language proce...

    2020

  21. [29]

    Generating soap notes from doctor-patient conversations using modular summarization techniques,

    K. Krishna, S. Khosla, J. P. Bigham, and Z. C. Lipton, “Generating soap notes from doctor-patient conversations using modular summarization techniques,” arXiv preprint arXiv:2005.01795 , 2020

    2005 arXiv

  22. [30]

    Generating more faithful and consistent soap notes using attribute-specific parameters,

    S. Ramprasad, E. Ferracane, and S. P. Selvaraj, “Generating more faithful and consistent soap notes using attribute-specific parameters,” in Machine Learning for Healthcare Conference . PMLR, 2023, pp. 631–649

    2023

  23. [31]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

    2019

  24. [32]

    Okapi at trec-3,

    S. E. Robertson, S. Walker, S. Jones, M. M. Hancock-Beaulieu, M. Gat- ford et al., “Okapi at trec-3,” Nist Special Publication Sp , vol. 109, p. 109, 1995

    1995

  25. [33]

    Dense passage retrieval for open-domain question answering,

    V . Karpukhin, B. O ˘guz, S. Min, P. Lewis, L. Wu, S. Edunov, D. Chen, and W.-t. Yih, “Dense passage retrieval for open-domain question answering,” arXiv preprint arXiv:2004.04906 , 2020

    2004 arXiv

  26. [34]

    Experimental approach toward training and analysing siamese deep neural network for sentence with no repeated expressions,

    H. Tripathi, “Experimental approach toward training and analysing siamese deep neural network for sentence with no repeated expressions,” in 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT) . IEEE, 2023, pp. 1–5

    2023

  27. [35]

    Reciprocal rank fusion outperforms condorcet and individual rank learning methods,

    G. V . Cormack, C. L. Clarke, and S. Buettcher, “Reciprocal rank fusion outperforms condorcet and individual rank learning methods,” in Proceedings of the 32nd international ACM SIGIR conference on Research and development in information retrieval , 2009, pp. 758–759

    2009

  28. [36]

    Lora: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “Lora: Low-rank adaptation of large language models,” arXiv preprint arXiv:2106.09685 , 2021

    2021 arXiv

  29. [37]

    How to fine-tune: Focus on effective datasets,

    Meta, “How to fine-tune: Focus on effective datasets,” https://ai.meta. com/blog/how-to-fine-tune-llms-peft-dataset-curation/, 2024

    2024

  30. [38]

    Stanford alpaca: An instruction-following llama model,

    R. Taori, I. Gulrajani, T. Zhang, Y . Dubois, X. Li, C. Guestrin, P. Liang, and T. B. Hashimoto, “Stanford alpaca: An instruction-following llama model,” https://github.com/tatsu-lab/stanford_alpaca, 2023

    2023

  31. [39]

    Rouge: A package for automatic evaluation of summaries,

    C.-Y . Lin, “Rouge: A package for automatic evaluation of summaries,” in Text summarization branches out , 2004, pp. 74–81

    2004

  32. [40]

    Bertscore: Evaluating text generation with bert,

    T. Zhang, V . Kishore, F. Wu, K. Q. Weinberger, and Y . Artzi, “Bertscore: Evaluating text generation with bert,” arXiv preprint arXiv:1904.09675, 2019

    1904 arXiv

  33. [41]

    Re-examining system-level corre- lations of automatic summarization evaluation metrics,

    D. Deutsch, R. Dror, and D. Roth, “Re-examining system-level corre- lations of automatic summarization evaluation metrics,” arXiv preprint arXiv:2204.10216, 2022

    2022 arXiv

  34. [42]

    Consultation checklists: Standardising the human evaluation of medical note generation,

    A. Savkov, F. Moramarco, A. P. Korfiatis, M. Perera, A. Belz, and E. Reiter, “Consultation checklists: Standardising the human evaluation of medical note generation,” arXiv preprint arXiv:2211.09455 , 2022

    2022 arXiv

  35. [43]

    Interrater reliability: the kappa statistic,

    M. L. McHugh, “Interrater reliability: the kappa statistic,” Biochemia medica, vol. 22, no. 3, pp. 276–282, 2012

    2012

Pith tools

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