Knowledge-Augmented Language Models Interpreting Structured Chest X-Ray Findings

Alexander Davis; Jia-Hao Lim; Rafael Souza

REVIEW 5 major objections 7 minor 1 cited by

Knowledge-Augmented Language Models Interpreting Structured Chest X-Ray Findings

T0 review · 5 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Structured text lets a 1B text-only LLM top multimodal X-ray models

desk verdict The text-centric idea is worth a look, but with the image-to-text pipeline undisclosed, the headline SOTA numbers are not trustworthy. read the letter →

arxiv 2505.01711 v1 pith:PLS2UGOK submitted 2025-05-03 cs.CV

Alexander Davis , Rafael Souza , Jia-Hao Lim This is my paper · ORCID

classification cs.CV

keywords chestX-rayinterpretationlargelanguagemodelsstructuredtextualrepresentationradiologyknowledgegraphreportgenerationvisualquestionansweringmultimodalfoundationinstructiontuning

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

CXR-TextInter claims that a text-only large language model, given a structured textual description of a chest X-ray instead of the image itself, can outperform larger multimodal foundation models on pathology detection, report generation, and visual question answering. The authors fine-tune a 1-billion-parameter LLM on the MediInstruct-CXR instruction dataset, add radiology knowledge-graph embeddings to every input token, and evaluate on the CXR-ClinEval benchmark. There the model reaches 47.1 Macro F1 over 14 pathologies, a report score of 0.848, and 0.537 VQA accuracy, ahead of CheXagent and Med-PaLM-M. A blinded panel of radiologists rated its outputs higher on accuracy, completeness, and clinical relevance and preferred them in 60 percent of cases. The paper's broader claim is that visual information can be separated from reasoning: structure the image into text, and text-centric LLMs can do medical image interpretation.

What carries the argument

The load-bearing component is the structured textual representation $T_{\text{img}}$: an upstream image-analysis pipeline $f_{\text{img2text}}$, not detailed in the paper, converts a chest X-ray into tokens that list findings, attributes, anatomical locations, and spatial relationships, and this text is the LLM's only view of the image. Around it, the method adds a Radiology Knowledge Graph $G=(V,E)$ whose concept embeddings are averaged into a sequence-level vector and added to every input token embedding, injecting medical context. The LLM is a decoder-only Transformer fine-tuned with cross-entropy on instruction-response pairs. The mechanism is that the structured text supplies the visual facts, the knowledge vector biases reasoning toward medically consistent conclusions, and the pretrained text-only model contributes its generative and reasoning abilities.

What would settle it

Take a set of CXRs with ground-truth findings, run the upstream pipeline, and check whether each pathology label appears in the resulting structured text; any pathology missing from the text should also be missing from CXR-TextInter's reports and VQA answers. A direct version of the test is to feed the model the ground-truth structured description from a radiologist's report instead of the pipeline output: if scores rise substantially, the pipeline is the binding constraint. Alternatively, delete one finding from $T_{\text{img}}$ and verify that the model's accuracy on questions about that finding collapses.

Watch

Extended reading notes

Core claim

The central claim is that a text-only LLM becomes the best CXR interpreter when the image is replaced by a rich structured text representation produced by an upstream vision pipeline. The paper argues that entities, attributes, anatomical locations, and spatial relations encoded as tokens are sufficient input for a fine-tuned LLM to detect pathologies, generate reports, and answer clinical questions, and that injecting a radiology knowledge graph improves performance most on rare and critical findings. The ablation, 47.1 versus 42.5 Macro F1 without the knowledge module, 0.848 versus 0.810 report score, and 0.537 versus 0.485 VQA accuracy, attributes a major share of the gain to that module. The authors also report a lower hallucination rate, 4.8 percent versus 8.1 percent for CheXagent, because the model reads explicitly detected entities rather than raw pixels. They present the results as evidence that decoupling visual analysis from language reasoning is a viable alternative to end-to-end multimodal models.

Load-bearing premise

The results stand on the unstated assumption that the undisclosed upstream image-to-text pipeline converts a chest X-ray into structured text that preserves every clinically relevant finding, location, and relationship; if that pipeline omits or misstates a finding, the LLM cannot recover it, so the reported scores partly measure the pipeline's detection ability rather than the LLM's interpretation.

Editorial extensions

If this is right

A 1B text-only LLM can outperform 3B-84B multimodal foundation models on CXR detection, report generation, and VQA when the image is pre-structured into text.
Adding medical knowledge-graph embeddings improves detection of rare pathologies and reduces medically inconsistent statements, as shown by the ablation and error analysis.
Hallucinated findings drop to 4.8 percent versus 8.1 percent for CheXagent because the LLM reads detected entities instead of raw pixels.
The CXR-ClinEval benchmark provides a reusable evaluation suite for LLM-centric medical image interpreters across classification, generation, and question answering.
Radiologists preferred CXR-TextInter outputs in 60 percent of blinded comparisons, suggesting the quality difference is clinically visible.

Reading between the lines

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

If the paradigm holds, the bottleneck shifts from model scale and multimodal fusion to the quality of image-to-text translation; improving the upstream pipeline should raise the ceiling directly, since the LLM cannot recover findings the text omits.
The same decoupling could transfer to ultrasound, mammography, or pathology slides whenever a reliable structured descriptor exists, letting medical imaging interpretation track text-LLM progress rather than multimodal training.
A direct test of the knowledge module's contribution would be to replace averaged knowledge embeddings with per-case retrieval of relevant concepts; if per-case retrieval further lifts rare-pathology F1, the current averaging under-uses the graph.
Combining structured image text with patient history and laboratory values is a natural next step, since both are text and the current design already accepts additional textual input alongside $T_{\text{img}}$.

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.

Desk Editor's Note

read the letter

Bottom line: this paper has a plausible idea and some honest acknowledgments, but it hides the component that actually does the visual work, so the central SOTA claim can't be assessed from the manuscript. The reader's rejection is right, and I'd desk-reject it too.

What's genuinely new: the proposal to decouple CXR interpretation into an upstream image-to-text stage and a text-only LLM stage, and the creation of a large instruction dataset and a multi-task benchmark to support that paradigm. The knowledge integration module (average of concept embeddings added to all token embeddings) is simple, but the ablation suggests it helps, at least on this benchmark. Those are legitimate contributions if the pipeline is made public and evaluation fixed.

Where it falls apart: Section 3.1 defines f_img2text and says 'not detailed herein.' That's the whole ballgame. The LLM only sees Timg, so the pathology detection F1 scores in Table 1 measure how well the LLM echoes labels that the pipeline already produced. The paper's own Section 4.8 admits missing findings are 'primarily dependent on the upstream image-to-text pipeline's detection capabilities'—so the LLM can't recover what isn't in the text. Comparing against multimodal baselines that read pixels is apples-to-oranges unless the pipeline's error profile is characterized.

The rest is consistent with that story: no error bars or significance tests anywhere, single-run ablation, three radiologists with no inter-rater agreement or confidence intervals, and a benchmark that isn't released. The citation list leans heavily on a particular group's papers, which looks like padding rather than engagement.

I'm not accusing the authors of anything. The conclusion and error analysis do mention the pipeline dependency as a limitation. But the abstract and introduction still claim SOTA without that caveat, and the evidence as written can't support it.

Who's it for? Researchers working on structured-input LLM methods might find the framework ideas worth a skim, but I wouldn't build on these numbers. It doesn't deserve referee time until the pipeline is disclosed and the evaluation is made fair.

Recommendation: desk reject as submitted, with an invitation to resubmit after releasing the pipeline and running a proper comparison.

Referee Report

5 major / 7 minor

Summary. The paper proposes CXR-TextInter, a framework that replaces direct image understanding with a two-stage pipeline: an upstream image analysis system f_img2text converts chest X-rays into structured text Timg (Section 3.1), and a text-only LLM, augmented by a learned knowledge-graph integration module, is fine-tuned to produce reports and answers from Timg. The authors introduce the MediInstruct-CXR training dataset and the CXR-ClinEval benchmark, and report results on pathology detection, report generation, VQA, differential diagnosis ranking, and a human preference study, claiming state-of-the-art performance over multimodal baselines such as CheXagent and Med-PaLM-M. The paper acknowledges in Sections 4.8 and 5 that performance is constrained by the upstream pipeline, but the pipeline itself is not described.

Significance. If the claims were fully supported, this would establish a useful alternative architectural paradigm for medical imaging AI, showing that strong text-only LLMs can interpret images once a sufficiently rich structured text representation is available. The manuscript has clear strengths: the LLM-side architecture is formally specified with equations for knowledge integration (Section 3.2), the evaluation includes several tasks and a blinded human study, and the authors are explicit about the dependency on the upstream representation. However, the central quantitative claim is unverifiable in the submitted form because the upstream f_img2text pipeline is undisclosed; regardless of its internal quality, the pathology detection numbers cannot be attributed to the LLM. The paper therefore does not currently support the abstract's claim of state-of-the-art CXR interpretation.

major comments (5)

[Section 3.1] The core of the method, f_img2text, is explicitly 'not detailed herein', yet it is the only component that converts pixels into the findings, locations, and relationships used in every evaluation task. Because Timg is the sole input to the LLM, the reported Macro F1/Micro F1 and VQA numbers in Tables 1-4 measure the combined pipeline-plus-LLM system, and if the pipeline already emits the pathology labels, the detection task is circular: the LLM would only need to echo Timg. The manuscript must specify the pipeline's architecture, training data, detection performance, and, crucially, demonstrate that Timg does not contain the test labels; without this, the SOTA claim in Table 1 cannot be assessed.
[Section 4.8] The error analysis states that missing findings are 'primarily dependent on the upstream image-to-text pipeline's detection capabilities.' This is a direct admission that the LLM cannot recover information absent from Timg, so any improvement in missing-finding rates cannot be attributed to the proposed framework. Similarly, the reduced hallucination rate in Table 6 is expected by construction, since the LLM only generates from a list of detected entities; it does not show that CXR-TextInter interprets images better than multimodal baselines. The paper should report the pipeline's standalone finding-detection performance and measure the information loss between image and Timg.
[Tables 1-6] All quantitative results are point estimates. There are no error bars, confidence intervals, significance tests, or multiple training seeds, and the ablation study (Table 2) compares single runs. Given the modest differences between models (e.g., Report Score 0.848 vs. 0.835 in Table 1), the claimed superiority of CXR-TextInter and the contribution of the knowledge module in Section 4.4 are not statistically supported. The authors should provide variance estimates and significance tests, or at least multiple seeds.
[Section 4.7] The human evaluation uses three radiologists on 100 CXRs but reports only average Likert scores and raw preference percentages. No inter-rater agreement (e.g., Cohen's kappa or Fleiss' kappa), confidence intervals, or significance tests are given, so the statement that radiologists 'significantly' prefer CXR-TextInter is unsupported. Report agreement and a proper test of preference (e.g., Wilcoxon signed-rank on per-case scores).
[Section 3.2] The knowledge integration module is a central contribution, but the mapping φ(S) that identifies relevant concepts from the input sequence, the construction of the Radiology Knowledge Graph G, and the choice of aggregation (Equation 4) are not specified. As a result, the ablation in Table 2 cannot be interpreted: the improvement could come from any number of implementation choices. The authors should define φ(S), state how G is built (including whether it uses the CXR-ClinEval or MediInstruct-CXR data), and report the retrieval/aggregation design.

minor comments (7)

[Section 1] The text introduces 'CXR-TextInter (placeholder name)', indicating the name is not final; remove this before publication.
[Tables 1-6] There are formatting errors in model names, e.g., 'LLaV A-Rad' and 'CXR-T extInter'; fix spacing.
[Section 4.2] A 'composite average score' is mentioned but never defined or reported; either define it or remove the reference.
[Section 4.5] The Differential Diagnosis Ranking task is called 'hypothetical' and MRR is reported without a description of the candidate set or ground truth; specify the task protocol.
[Section 3.4] The decoding strategy used in the reported experiments is not stated (greedy, beam, or sampling); specify it.
[References] References [27] and [29] are the same paper (Singhal et al.), and several baselines (MARIA, LLaVA-Rad) are described without citations; add the appropriate references.
[Datasets] The paper does not state whether MediInstruct-CXR and CXR-ClinEval will be released; for reproducibility, release details are needed.

Circularity Check

2 steps flagged · score 8.0 of 10

Pathology-detection, VQA, and report-generation results reduce to the undisclosed upstream pipeline's labels, because Timg already contains the findings being predicted.

self definitional [Section 3.1 (Structured Textual Representation and Input Formatting); Section 4.2 (Evaluation Tasks and Metrics)]
""This representation is designed to capture crucial visual information in a format amenable to LLM processing. It goes beyond simple captions, encoding specific entities (e.g., 'consolidation,' 'effusion'), their attributes (e.g., 'size,' 'opacity'), precise anatomical locations (e.g., 'right lower lobe,' 'left costophrenic angle'), and spatial relationships between entities" (Sec. 3.1); evaluation includes "multi-label pathology detection for 14 common thoracic pathologies" (Sec. 4.2)."

The pathology-detection task's target labels are the same 14 findings that Timg explicitly enumerates. Since Timg is the sole image-derived input to the LLM (Eq. 1), the model can satisfy the task by copying the listed entities into its response; any finding not listed in Timg cannot be reported, as Section 4.8 admits. Therefore the Macro/Micro F1 scores in Table 1 measure the upstream pipeline's pre-extracted labels plus text formatting, not image interpretation by the LLM, making the SOTA comparison circular.
fitted input called prediction [Section 3.1; Section 4.8 (Error Analysis)]
""The raw CXR image I is processed by a separate, sophisticated image analysis pipeline fimg2text (not detailed herein), which generates a structured textual representation Timg" (Sec. 3.1); "the rate of missing findings is comparable (primarily dependent on the upstream image-to-text pipeline's detection capabilities)" (Sec. 4.8)."

The paper treats fimg2text as an external component, yet the evaluation's headline quantity (pathology detection) is exactly what this fitted, undisclosed pipeline produces. Feeding those extracted findings back as input and scoring the LLM's output against the same finding labels is a fitted input renamed as a prediction: the reported F1 is bounded by the pipeline's detector, and the LLM adds no visual evidence. The paper's own error analysis confirms this by attributing missing-finding errors to the pipeline rather than to the model.

full rationale

The paper's core quantitative claim is circular for the tasks that support it. Section 3.1 defines Timg as encoding specific findings, attributes, and locations, and Section 4.2 evaluates multi-label pathology detection for 14 thoracic pathologies, i.e., the same entities. Because the LLM receives Timg as its sole visual input, its F1 scores measure whether the model repeats the upstream pipeline's pre-extracted findings, not whether it detects pathology from pixels. The paper does not describe f_img2text or its training data, so the comparison against CheXagent, Med-PaLM-M, GPT-4V, and the other multimodal baselines is not an apples-to-apples test of visual understanding. Section 4.8 confirms the ceiling: missing-finding errors are 'primarily dependent on the upstream image-to-text pipeline's detection capabilities,' so any finding absent from Timg cannot be recovered by the LLM. Report generation and VQA similarly restate Timg content, making the aggregate SOTA claim hostage to the hidden pipeline. There is no self-citation chain or imported uniqueness theorem; the circularity is in the evaluation design, not in the references. Consequently, the quantitative contribution cannot be regarded as an independent demonstration of LLM-based CXR interpretation.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The paper's central claim depends on an undisclosed image-to-text pipeline, an unspecified knowledge graph, and an unreleased dataset/benchmark. The learned concept embeddings and projection parameters are fitted on the authors' own data, and the evaluation offers no external validation.

free parameters (3)

Knowledge graph concept embeddings k_c = Learned during fine-tuning on MediInstruct-CXR
Eq. 3-5: each concept c in G has a learned embedding; these are jointly optimized with the LLM and determine the knowledge vector added to all token embeddings.
Projection matrix W_proj,k and bias b_proj,k = Learned during fine-tuning
Eq. 5: maps the averaged knowledge vector into the model embedding dimension; central to the claimed contribution of the knowledge module.
LLM fine-tuning parameters (base model and training hyperparameters) = Undisclosed; model size 1B
Section 3.3 gives only the loss and optimizer names, not learning rate, schedule, epochs, or data order; the central results depend on these choices.

assumptions (3)

domain assumption The upstream image-to-text pipeline f_img2text reliably converts a CXR into a structured text representation that preserves all clinically relevant findings, attributes, locations, and relationships.
Section 3.1: the pipeline is explicitly "not detailed herein", yet it is the sole visual input to the LLM. If this assumption fails, the entire framework's interpretation ability is untested.
domain assumption The Radiology Knowledge Graph G is a complete and correct source of medical concepts and relations, and the mapping phi(S) selects the relevant concepts for a given input.
Section 3.2: no source, size, or construction of G is given, and phi(S) is an unspecified function. The ablation claims a 4.6-point F1 gain from this module.
domain assumption The CXR-ClinEval benchmark and MediInstruct-CXR dataset are correctly labeled, representative of clinical practice, and free of train/test leakage.
Section 4.2: dataset construction, size, and split methodology are not described; all quantitative claims are evaluated on this self-created benchmark.

invented entities (1)

Radiology Knowledge Graph G = (V, E)
purpose: Supplies medical concept embeddings that are averaged and added to every token embedding (Eq. 4-5) to improve clinical reasoning.
The graph is introduced without a source, schema, or validation set; its completeness and correctness are assumed, and the claimed benefit rests entirely on this entity.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Knowledge-Augmented Language Models Interpreting Structured Chest X-Ray Findings." pith.science (2026). https://pith.science/paper/PLS2UGOK

@misc{pith2026250501711,
  author       = {Pith},
  title        = {Pith review of: Knowledge-Augmented Language Models Interpreting Structured Chest X-Ray Findings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PLS2UGOK}},
  note         = {Machine review of arXiv:2505.01711}
}

read the original abstract

Automated interpretation of chest X-rays (CXR) is a critical task with the potential to significantly improve clinical workflow and patient care. While recent advances in multimodal foundation models have shown promise, effectively leveraging the full power of large language models (LLMs) for this visual task remains an underexplored area. This paper introduces CXR-TextInter, a novel framework that repurposes powerful text-centric LLMs for CXR interpretation by operating solely on a rich, structured textual representation of the image content, generated by an upstream image analysis pipeline. We augment this LLM-centric approach with an integrated medical knowledge module to enhance clinical reasoning. To facilitate training and evaluation, we developed the MediInstruct-CXR dataset, containing structured image representations paired with diverse, clinically relevant instruction-response examples, and the CXR-ClinEval benchmark for comprehensive assessment across various interpretation tasks. Extensive experiments on CXR-ClinEval demonstrate that CXR-TextInter achieves state-of-the-art quantitative performance across pathology detection, report generation, and visual question answering, surpassing existing multimodal foundation models. Ablation studies confirm the critical contribution of the knowledge integration module. Furthermore, blinded human evaluation by board-certified radiologists shows a significant preference for the clinical quality of outputs generated by CXR-TextInter. Our work validates an alternative paradigm for medical image AI, showcasing the potential of harnessing advanced LLM capabilities when visual information is effectively structured and domain knowledge is integrated.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

Toward a Vision-Language Foundation Model for Medical Data: Multimodal Dataset and Benchmarks for Vietnamese PET/CT Report Generation
cs.CV 2025-09 conditional novelty 6.0 of 10

A new Vietnamese PET/CT-report dataset improves medical VLM report generation and VQA, but clinical F1 scores remain modest.

Reference graph

Works this paper leans on

33 extracted references · 20 canonical work pages · cited by 1 Pith paper

[1]

Davis et al

Rajpurkar, P., Irvin, J., Ball, R.L., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C.P., et al.: Deep learning for chest radiograph diagnosis: 14 A. Davis et al. A retrospective comparison of the chexnext algorithm to practicing radiologists. PLoS medicine 15(11), e1002686

work page
[2]

arXiv preprint arXiv:1711.05225

Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C., Shpanskaya, K., et al.: Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225

arXiv
[3]

Artificial Intelligence in Medicine106, 101878

Monshi, M.M.A., Poon, J., Chung, V.: Deep learning in generating radiology re- ports: A survey. Artificial Intelligence in Medicine106, 101878

work page
[4]

arXiv preprint arXiv:2310.18689

Azad, B., Azad, R., Eskandari, S., Bozorgpour, A., Kazerouni, A., Rekik, I., Mer- hof, D.: Foundational models in medical imaging: A comprehensive survey and future vision. arXiv preprint arXiv:2310.18689

arXiv
[5]

arXiv preprint arXiv:2401.12208

Chen, Z., Varma, M., Delbrouck, J.B., Paschali, M., Blankemeier, L., Van Veen, D., Valanarasu, J.M.J., Youssef, A., Cohen, J.P., Reis, E.P., et al.: Chexagent: Towards a foundation model for chest x-ray interpretation. arXiv preprint arXiv:2401.12208

arXiv
[6]

Nejm Ai 1(3), AIoa2300138

Tu, T., Azizi, S., Driess, D., Schaekermann, M., Amin, M., Chang, P.C., Carroll, A., Lau, C., Tanno, R., Ktena, I., et al.: Towards generalist biomedical ai. Nejm Ai 1(3), AIoa2300138

work page
[7]

In: CLEF (working notes)

Shi, L., Liu, F., Rosen, M.P.: Deep multimodal learning for medical visual question answering. In: CLEF (working notes)

work page
[8]

In: Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics

Zhou, Y., Long, G.: Multimodal event transformer for image-guided story ending generation. In: Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics. pp. 3434–3444 (2023)

work page 2023

Show all 33 references

[9]

In: Findings of the Association for Computational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11-16, 2024

Zhou, Y., Li, X., Wang, Q., Shen, J.: Visual in-context learning for large vision- language models. In: Findings of the Association for Computational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11-16, 2024. pp. 15890– 15902. Association for Computation...

2024
[10]

Zhou, Y., Zhang, J., Chen, G., Shen, J., Cheng, Y.: Less is more: Vision repre- sentation compression for efficient video generation with large language models (2024)

2024
[11]

arXiv preprint arXiv:2410.19732 (2024)

Zhou, Y., Rao, Z., Wan, J., Shen, J.: Rethinking visual dependency in long-context reasoning for large vision-language models. arXiv preprint arXiv:2410.19732 (2024)

2024 arXiv
[12]

arXiv preprint arXiv:2501.15167 (2025)

He, Y., Wang, J., Li, K., Wang, Y., Sun, L., Yin, J., Zhang, M., Wang, X.: En- hancing intent understanding for ambiguous prompts through human-machine co- adaptation. arXiv preprint arXiv:2501.15167 (2025)

2025
[13]

In: The Thirteenth International Confer- ence on Learning Representations (2025), https://openreview.net/forum?id= N1vYivuSKq

Zhou, Y., Shen, J., Cheng, Y.: Weak to strong generalization for large lan- guage models with multi-capabilities. In: The Thirteenth International Confer- ence on Learning Representations (2025), https://openreview.net/forum?id= N1vYivuSKq

2025
[14]

arXiv preprint arXiv:2503.23512 (2025)

Yi, Q., He, Y., Wang, J., Song, X., Qian, S., Zhang, M., Sun, L., Shi, T.: Score: Story coherence and retrieval enhancement for ai narratives. arXiv preprint arXiv:2503.23512 (2025)

2025
[15]

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021 (2021)

Zhou, Y., Geng, X., Shen, T., Pei, J., Zhang, W., Jiang, D.: Modeling event- pair relations in external knowledge graphs for script reasoning. Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021 (2021)

2021
[16]

In: Proceedings of the ACM Web Conference 2022

Zhou, Y., Geng, X., Shen, T., Long, G., Jiang, D.: Eventbert: A pre-trained model for event correlation reasoning. In: Proceedings of the ACM Web Conference 2022. pp. 850–859 (2022)

2022
[17]

Advances in neural information pro- cessing systems 30 CXR-TextInter 15

Vaswani,A.,Shazeer,N.,Parmar,N.,Uszkoreit,J.,Jones,L.,Gomez,A.N.,Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in neural information pro- cessing systems 30 CXR-TextInter 15
[18]

In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers)

Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: Pre-training of deep bidi- rectional transformers for language understanding. In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologi...

2019
[19]

Journal of machine learning research21(140), 1–67

Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., Liu, P.J.: Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research21(140), 1–67
[20]

OpenAI blog1(8), 9

Radford,A.,Wu,J.,Child,R.,Luan,D.,Amodei,D.,Sutskever,I.,etal.:Language models are unsupervised multitask learners. OpenAI blog1(8), 9
[21]

Advances in neural information processing systems33, 1877–1901

Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J.D., Dhariwal, P., Nee- lakantan, A., Shyam, P., Sastry, G., Askell, A., et al.: Language models are few-shot learners. Advances in neural information processing systems33, 1877–1901

1901
[22]

arXiv preprint arXiv:2203.15556

Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., Casas, D.d.L., Hendricks, L.A., Welbl, J., Clark, A., et al.: Training compute- optimal large language models. arXiv preprint arXiv:2203.15556

arXiv
[23]

arXiv preprint arXiv:2302.13971

Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.A., Lacroix, T., Rozière, B., Goyal, N., Hambro, E., Azhar, F., et al.: Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971

arXiv
[24]

arXiv preprint arXiv:2303.18223 1(2)

Zhao, W.X., Zhou, K., Li, J., Tang, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., et al.: A survey of large language models. arXiv preprint arXiv:2303.18223 1(2)

arXiv
[25]

Briefings in bioinformatics 23(6), bbac409

Luo,R.,Sun,L.,Xia,Y.,Qin,T.,Zhang,S.,Poon,H.,Liu,T.Y.:Biogpt:generative pre-trained transformer for biomedical text generation and mining. Briefings in bioinformatics 23(6), bbac409
[26]

Digital Health 10, 20552076241308987

Abbas, A., Lee, M., Shanavas, N., Kovatchev, V.: Clinical concept annotation with contextual word embedding in active transfer learning environment. Digital Health 10, 20552076241308987
[28]

arXiv preprint arXiv:2303.13375

Nori, H., King, N., McKinney, S.M., Carignan, D., Horvitz, E.: Capabilities of gpt-4 on medical challenge problems. arXiv preprint arXiv:2303.13375

arXiv
[29]

Nature Medicine pp

Singhal, K., Tu, T., Gottweis, J., Sayres, R., Wulczyn, E., Amin, M., Hou, L., Clark, K., Pfohl, S.R., Cole-Lewis, H., et al.: Toward expert-level medical question answering with large language models. Nature Medicine pp. 1–8
[30]

Information Fusion p

He, K., Mao, R., Lin, Q., Ruan, Y., Lan, X., Feng, M., Cambria, E.: A survey of large language models for healthcare: from data, technology, and applications to accountability and ethics. Information Fusion p. 102963
[31]

arXiv preprint arXiv:2406.03712

Liu, L., Yang, X., Lei, J., Liu, X., Shen, Y., Zhang, Z., Wei, P., Gu, J., Chu, Z., Qin, Z., et al.: A survey on medical large language models: Technology, application, trustworthiness, and future directions. arXiv preprint arXiv:2406.03712

arXiv
[32]

Cureus15(5)

Karabacak, M., Margetis, K.: Embracing large language models for medical appli- cations: opportunities and challenges. Cureus15(5)
[33]

Nature medicine 30(9), 2613–2622

Hager, P., Jungmann, F., Holland, R., Bhagat, K., Hubrecht, I., Knauer, M., Viel- hauer, J., Makowski, M., Braren, R., Kaissis, G., et al.: Evaluation and mitigation of the limitations of large language models in clinical decision-making. Nature medicine 30(9), 2613–2622
[34]

Applied Clinical Informatics15(02), 306–312

Spotnitz, M., Idnay, B., Gordon, E.R., Shyu, R., Zhang, G., Liu, C., Cimino, J.J., Weng, C.: A survey of clinicians’ views of the utility of large language models. Applied Clinical Informatics15(02), 306–312

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Davis et al

Rajpurkar, P., Irvin, J., Ball, R.L., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C.P., et al.: Deep learning for chest radiograph diagnosis: 14 A. Davis et al. A retrospective comparison of the chexnext algorithm to practicing radiologists. PLoS medicine 15(11), e1002686

work page

[2] [2]

arXiv preprint arXiv:1711.05225

Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C., Shpanskaya, K., et al.: Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225

arXiv

[3] [3]

Artificial Intelligence in Medicine106, 101878

Monshi, M.M.A., Poon, J., Chung, V.: Deep learning in generating radiology re- ports: A survey. Artificial Intelligence in Medicine106, 101878

work page

[4] [4]

arXiv preprint arXiv:2310.18689

Azad, B., Azad, R., Eskandari, S., Bozorgpour, A., Kazerouni, A., Rekik, I., Mer- hof, D.: Foundational models in medical imaging: A comprehensive survey and future vision. arXiv preprint arXiv:2310.18689

arXiv

[5] [5]

arXiv preprint arXiv:2401.12208

Chen, Z., Varma, M., Delbrouck, J.B., Paschali, M., Blankemeier, L., Van Veen, D., Valanarasu, J.M.J., Youssef, A., Cohen, J.P., Reis, E.P., et al.: Chexagent: Towards a foundation model for chest x-ray interpretation. arXiv preprint arXiv:2401.12208

arXiv

[6] [6]

Nejm Ai 1(3), AIoa2300138

Tu, T., Azizi, S., Driess, D., Schaekermann, M., Amin, M., Chang, P.C., Carroll, A., Lau, C., Tanno, R., Ktena, I., et al.: Towards generalist biomedical ai. Nejm Ai 1(3), AIoa2300138

work page

[7] [7]

In: CLEF (working notes)

Shi, L., Liu, F., Rosen, M.P.: Deep multimodal learning for medical visual question answering. In: CLEF (working notes)

work page

[8] [8]

In: Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics

Zhou, Y., Long, G.: Multimodal event transformer for image-guided story ending generation. In: Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics. pp. 3434–3444 (2023)

work page 2023