Pith. sign in

REVIEW 5 major objections 6 minor 49 references

Linguistically-Aligned and Visually-Grounded Preference Optimization for Clinically-Augmented Medical Report Generation

T0 review · 5 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read DPO-Clin shows that medical report generation improves when preference optimization isolates clinical findings and grounds them in image evidence.

desk verdict Solid incremental MRG paper: DPO-Clin's entity-focused preference pairs and visual-context inversion show consistent gains on independent CheXbert metrics, but its RaTEScore gains are partly circular because the metric and the method share the same NER, so the clinical-factuality claim needs external validation. read the letter →

arxiv 2608.08494 v2 pith:54UREK34 submitted 2026-08-09 cs.CV

Qiang Hu , Yuxuan Luo , Yingjie Guo , Hao Wang , Qimei Wang , Qiang Li , Zhiwei Wang This is my paper · ORCID
classification cs.CV
keywords medicalreportgenerationdirectpreferenceoptimizationclinicalfactualityvision-languagealignmententity-leveldiagnosislatentriskmitigationchestX-rayendoscopy
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

This paper claims that the standard way of applying Direct Preference Optimization (DPO) to medical report generation—pairing a model's raw output against the ground-truth report—dilutes training with clinically irrelevant wording differences. It proposes DPO-Clin, a post-training framework that first diagnoses reports at the level of medical entities and their presence/absence assertions, then builds preference pairs that differ only in those clinical facts. It further adds a multi-modal objective that flips the textual preference when the input image switches to a visually different but textually matching one, forcing the model to ground assertions in image evidence. On MIMIC-CXR, IU X-Ray, and an in-house endoscopy dataset, the framework improves clinical metrics over the supervised fine-tuning baselines and over prior DPO-based report-generation methods. If correct, this redirects preference optimization for medical text from surface similarity toward factual and visually grounded content.

What carries the argument

The load-bearing object is the linguistically-aligned preference pair generated by the ECD module, together with the M2DPO objective. ECD uses RaTE-NER to parse each report into entities with semantic embeddings and assertion statuses, solves a threshold-constrained linear assignment problem to match predicted entities to ground-truth entities, and classifies discrepancies into correct matches, extraneous entities, missing entities, and false assertions; a prompt built from those diagnoses drives GPT-4o to rewrite the predicted report so that it preserves the prediction's phrasing but states the ground truth's clinical facts. The M2DPO objective is the sum of two DPO terms, $\mathcal{L}_{\mathrm{M2DPO}}(x_1,x_2,y_1,y_2)=\mathcal{L}_{\mathrm{DPO}}(x_1,y_1,y_2)+\mathcal{L}_{\mathrm{DPO}}(x_2,y_2,y_1)$, which forces the model to prefer $y_1$ over $y_2$ under image $x_1$ and to reverse that preference under image $x_2$. Counterfactual latent-risk data is produced by replacing high-entropy correct entities with the second-most-probable candidate, subject to RadGraph-schema verification.

What would settle it

Select a random sample of paired predicted and ground-truth reports from MIMIC-CXR and the endoscopy set, have a clinician independently mark the medical findings and their present/absent status, and compare against the ECD module's diagnosis; if the module misses or flips a sizable fraction of clinically significant findings, then the supposedly linguistically-aligned preference reports are not actually isolating clinical discrepancies. The same human-annotated entities can replace ECD's entity set in a rerun of DPO-Clin: if the clinical-metric gains disappear, the framework's stated mechanism fails.

Watch

Extended reading notes

Core claim

The central discovery claim is that preference optimization for medical report generation can be made clinically effective by controlling what the preference comparison is allowed to differ on. The paper asserts that DPO-Clin, built on three mechanisms, significantly improves the SFT baselines on clinical-aware metrics and outperforms existing DPO-based MRG methods across two chest X-ray datasets and an endoscopy dataset. The explicit-error branch uses the Entity-level Clinical Diagnostic (ECD) module to extract medical entities and assertions, match them across predicted and ground-truth reports, and prompt an LLM to produce a preferred report that keeps the prediction's wording but adopts the ground truth's clinical facts; the multi-modal branch retrieves an image matching the predicted report and applies M2DPO, which inverts the textual preference when the visual context switches; the latent-risk branch identifies correct but uncertain entities and counterfactually swaps them for plausible alternatives. The reported result is that combining these branches lifts clinical efficacy metrics (14-label F1, RadGraph, RaTEScore, RadCliQ) on MIMIC-CXR and IU X-Ray, and the 2F1 normal/abnormal score on endoscopy, while improving or matching NLG metrics.

Load-bearing premise

The load-bearing premise is that the software module that identifies medical findings and whether they are described as present or absent is accurate enough to serve as ground truth, so any finding it misses or mislabels corrupts every preference pair, retrieved image, and counterfactual report built from that diagnosis.

Editorial extensions

If this is right

  • Preference pairs for clinical text should be constructed to isolate clinical discrepancies, rather than by directly contrasting model output with ground-truth text.
  • Adding a visual-context-triggered preference inversion (M2DPO) improves grounding of textual assertions in image evidence, as shown by more lesion-aligned cross-modal attention.
  • Explicitly optimizing against counterfactual, high-uncertainty correct predictions shifts the model's uncertainty distribution downward (mean 0.514 to 0.282 on MIMIC-CXR) and increases the count of correct entities.
  • The gains hold across two baseline architectures (R2GenGPT and RADAR) and across radiology and endoscopy, so the framework is portable to different MRG models and imaging modalities.
  • The dynamic masking of invalid instances is necessary: without it, joint training harms performance; replacing the linguistically-aligned preferred report with the raw ground-truth report degrades clinical metrics.

Reading between the lines

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

  • Beyond the paper: the ECD diagnosis—entity matching plus assertion comparison—could be reused as a standalone explainable factual-error annotation tool for auditing any MRG model, independent of the preference-optimization loop.
  • Beyond the paper: because the method's quality depends on the entity extractor, an immediate testable extension is to swap RaTE-NER for a newer or domain-specific extractor and measure whether the clinical-metric gains scale with extractor accuracy; the paper does not report this sensitivity.
  • Beyond the paper: the same linguistically-aligned preference construction could be applied to other clinical generation tasks, such as discharge summaries or procedure notes, wherever an entity-and-assertion diagnosis is available.
  • Beyond the paper: if GPT-4o's rewrites occasionally introduce subtle hallucinations not caught by ECD, the claimed isolation of clinical discrepancies is only as clean as the diagnosis; measuring the rewrite error rate on a human-labeled subset would settle this.
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

5 major / 6 minor

Summary. The manuscript proposes DPO-Clin, a post-training framework for medical report generation that aims to make Direct Preference Optimization focus on clinically relevant content and cross-modal grounding. The method has three components: an Entity-level Clinical Diagnostic (ECD) module built on RaTE-NER that diagnoses entity-level errors and guides GPT-4o to produce linguistically aligned preferred reports; M2DPO, a retrieval-augmented multimodal DPO variant that inverts textual preference when the visual context changes; and a counterfactual data construction step that targets correct-but-uncertain entity predictions. Experiments are reported on MIMIC-CXR, IU X-Ray, and a private in-house endoscopy dataset, using R2GenGPT and RADAR as baselines, with NLG, CheXbert-based clinical efficacy, and radiology-report-exclusive metrics. The paper reports consistent improvements over SFT baselines and existing DPO-based MRG methods, and ablations indicate that each component contributes.

Significance. If the results hold, the paper makes a useful contribution by redirecting DPO-based MRG from surface-level text matching toward entity-level clinical factuality and explicit visual grounding. The method is clearly specified, and the core idea of constructing linguistically aligned preference pairs under entity-level diagnosis is well motivated. The paper deserves credit for evaluating across two architectures and three datasets, and for including an informative ablation (Table 3) that isolates the contributions of the ECD-generated preferred report, the counterfactual latent-risk data, and the masking mechanism. A notable strength is that the CheXbert-based CE metrics are independent of the ECD module, so the central clinical gain is not entirely circular. However, the absence of any statistical significance testing, the sharing of RaTE-NER between the preference-construction pipeline and the RaTEScore evaluation metric, and the lack of verifiable details about the private endoscopy dataset and its NER fine-tuning currently prevent the paper from fully supporting its 'significantly improves' and 'robust generalizability' claims.

major comments (5)
  1. [Section 5.3, Tables 1-4] The central claim that DPO-Clin 'significantly improves' SFT baselines is not supported by any variance estimate, confidence interval, or hypothesis test. All tables report single point estimates, and several reported differences are small (e.g., MIMIC-CXR R2GenGPT B-1 0.411 vs. 0.423, or RADAR 14Ma-F1 0.460 vs. 0.490). The authors should provide standard deviations across multiple seeds or bootstrapped confidence intervals over report-level scores, and run paired significance tests at least for the CheXbert-based CE metrics that carry the main claim.
  2. [Sections 4.2 and 5.1] The ECD module uses RaTE-NER for entity extraction and assertion classification, and the RaTEScore evaluation metric uses the same RaTE-NER model (reference [47]). The RaTEScore improvements in Tables 1 and 3 are therefore partially circular: the model is trained to agree with the very entity/assertion judgments that the metric rewards. The independent CheXbert-based CE metrics partially mitigate this concern, but the manuscript should either relegate RaTEScore to a secondary role or provide an independent validation of ECD's entity/assertion accuracy on a held-out annotated set to show that the training signal corresponds to verifiable clinical factuality.
  3. [Sections 5.1 and 5.2] The in-house endoscopy dataset is private, and the fine-tuning of RaTE-NER is described only as using 'data annotated by endoscopists'; the paper provides no annotation protocol, no entity-level accuracy, and no dataset statistics. Because the endoscopy results are central to the claim of cross-modality generalization, the authors need to report at least entity-level NER performance on this domain, the size and annotation guidelines of the fine-tuning set, and ideally an external validation plan. Without this information, the endoscopy result is not verifiable and the generality claim is weakened.
  4. [Section 4.3] The two-stage retrieval protocol claims that the candidate pool guarantees 'absolute clinical equivalence' between the retrieved image x̂pre and the predicted report ypre, based on empty ECD error subsets. If RaTE-NER misses or mislabels an entity, the retrieved image may in fact be clinically incongruent, and the M2DPO inversion would then train the model on false preference signals. The paper should report retrieval pool sizes, hit rates, and an analysis of retrieval failures, and ideally validate a sample of retrieved quadruplets through manual review or an independent metric.
  5. [Section 4.4] The counterfactual report yunc is constructed by replacing an uncertain entity with the second most probable token and verifying the resulting sentence with RadGraph, but no human or clinical validation is provided that these sentences are grammatically and semantically plausible medical statements, and the frequency of knowledge-graph rejection is not reported. Since the latent-risk component contributes to the final gains in Table 3, the authors should show example counterfactuals and report the acceptance rate, or otherwise demonstrate that this data does not introduce spurious training signals.
minor comments (6)
  1. [Section 5.1] The '2F1' metric used for the endoscopy dataset is undefined; please specify how binary normal/abnormal labels are derived and how the F1 score is computed.
  2. [Equation (2)] The notation 'Σ_j M_{i,j} = {0,1}' is non-standard; it should be written as 'Σ_j M_{i,j} ∈ {0,1}' or stated explicitly that the sum is either 0 or 1.
  3. [Equation (4)] The conditioning context T_{<k} is not defined; please state that it denotes the token sequence before token k during decoding.
  4. [Section 4.2] The comparison of assertion statuses A_i = A_j is not specified; please clarify whether assertions are compared as exact strings or via a structured status mapping.
  5. [Table 3] The ablation row 'yGT → yGT' uses the same symbol yGT for both the raw ground-truth report and the LLM-generated linguistically aligned report, which is confusing; consider using distinct notation such as y_GT^+ for the generated preferred report.
  6. [Section 5.2] The manuscript does not state whether code, trained models, or the in-house dataset will be made available; a reproducibility statement would be helpful.

Circularity Check

1 steps flagged · score 4.0 of 10

RaTEScore gains are partially circular because the ECD preference construction and the RaTEScore metric share the same RaTE-NER entity/assertion model, though independent CheXbert and RadGraph metrics provide non-circular support.

  1. self definitional [Section 4.2 (ECD module) and Section 5.1 (Evaluation Metrics)]
    "ECD employs the RaTE-NER model [47] to parse both the predicted report ypre and the GT report yGT, extracting medical entities e along with their corresponding semantic embeddings E and assertion statuses A ∈ {Present, Absent} ... we employ three radiology report generation (RRG)-exclusive metrics (RadGraph [16], RadCliQ [42], RaTEScore [47])"

    The same reference [47] (RaTEScore) supplies the NER model that defines the ECD error subsets (P_EE, P_ME, P_FA) used to construct the DPO preferences, and it also supplies the RaTEScore metric used to evaluate clinical factuality. Thus, the reported RaTEScore improvements reflect in part optimization toward the very entity/assertion judgments that the metric rewards. This is a shared operational definition rather than independent verification. The circularity is partial because CheXbert-based CE, RadGraph, and RadCliQ are separate tools and also improve.

full rationale

No derivation in the paper is equivalent to its inputs by explicit equation, and the central DPO objective L_total is a standard preference loss applied to newly constructed pairs. The main non-independence is that ECD and RaTEScore both rely on RaTE-NER: the preference pairs are curated from RaTE-NER's entity/assertion diagnosis, and the headline RRG-exclusive metric computes entity-level scores with the same model. This makes the RaTEScore comparison partly self-referential but not fully circular, because results on CheXbert CE metrics, RadGraph, and RadCliQ are computed by independent tools and also improve. The endoscopy experiment is evaluated with a binary 2F1 CE metric rather than RaTEScore, so it is not subject to the same entanglement. Overall, the paper contains one partial circularity in the RaTEScore evaluation/construction loop, warranting a score of 4, not a higher score.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The central claim rests on the quality of the preference data. DPO-Clin assumes the NER-based ECD is an accurate map of clinical factuality, that GPT-4o transforms facts faithfully into the predicted report's style, that retrieval can find images 'absolutely clinically equivalent' to a report, and that token-level entropy is a valid uncertainty proxy. These are unverified domain assumptions, not derived results. The main tuned hyperparameters are tau, theta, and beta; no sensitivity analysis is reported. No new physical or ontological entities are introduced.

free parameters (4)
  • matching threshold tau = 0.4 for CXR, 0.5 for endoscopy
    Controls entity matching in ECD; chosen by hand, no sensitivity analysis.
  • uncertainty threshold theta = 0.5 on both CXR and endoscopy
    Threshold above which a correct entity is flagged as uncertain; selected empirically, not justified.
  • DPO temperature beta = 0.1
    Standard DPO hyperparameter; fixed, no sweep reported.
  • LoRA rank and alpha = not stated
    LoRA is applied to linear layers but rank and alpha are omitted, which is required for replication.
assumptions (7)
  • domain assumption RaTE-NER provides accurate medical entity extraction and assertion classification for both CXR and endoscopy reports.
    ECD's entire error typing (P_CM, P_EE, P_ME, P_FA) in Section 4.2 relies on the NER output; errors in NER propagate into preference pair quality and retrieval.
  • domain assumption GPT-4o, when given ECD diagnostic prompts, produces a preferred report that preserves the prediction's linguistic style while strictly adopting GT clinical facts.
    No verification step for the LLM-generated yGT is described; if GPT-4o introduces new entities or assertion errors, the preference labels are corrupted.
  • ad hoc to paper An image can be retrieved from the training set that is 'absolutely clinically equivalent' to a given report if ECD finds empty error subsets.
    Section 4.3 first-stage retrieval treats ECD's entity and assertion match as full clinical equivalence, an assumption about the completeness of the NER schema.
  • domain assumption CheXbert label conversion provides a valid ground truth for the 14 clinical-efficiency labels.
    Section 5.1 computes CE metrics using CheXbert; any CheXbert label errors affect the headline clinical gains.
  • ad hoc to paper Replacing an uncertain entity with the second most probable token, plus RadGraph edge verification, yields a clinically plausible counterfactual report.
    Section 4.4 uses this construction to create yunc; plausibility is checked only against RadGraph relations, not clinician review.
  • standard math Standard DPO assumptions hold for paired and inverted preference quadruplets.
    M2DPO applies DPO's implicit reward formulation in both directions (Eq. 3), inheriting DPO's assumptions about preference transitivity and reference coverage.
  • domain assumption The final model after DPO remains well-calibrated such that token entropy reflects clinical uncertainty.
    Latent risk mitigation assumes max token entropy of the reference policy is a meaningful proxy for clinical uncertainty (Eq. 4).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Linguistically-Aligned and Visually-Grounded Preference Optimization for Clinically-Augmented Medical Report Generation." pith.science (2026). https://pith.science/paper/54UREK34

@misc{pith2026260808494,
  author       = {Pith},
  title        = {Pith review of: Linguistically-Aligned and Visually-Grounded Preference Optimization for Clinically-Augmented Medical Report Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/54UREK34}},
  note         = {Machine review of arXiv:2608.08494}
}
read the original abstract

Despite significant advances in Medical Report Generation (MRG), the reliability remains constrained by the prevalence of factual errors. While Direct Preference Optimization (DPO) has emerged as a promising post-training paradigm to enhance the performance of Supervised Fine-Tuned (SFT) MRG models, existing DPO-based MRG methods typically adopt a naive preference construction that directly pairs model-generated reports with ground truth reports. This strategy inadvertently entangles critical clinical findings with clinically irrelevant linguistic characteristics, and fundamentally lacks explicit vision-language alignment. To address these challenges, we propose DPO-Clin, a novel post-training framework that focuses preference optimization on clinical findings and cross-modal alignment. First, we introduce the Entity-level Clinical Diagnostic (ECD) module to perform a precise entity-level factual diagnosis. ECD guides the generation of linguistically-aligned report preference pairs, isolating clinical discrepancies from linguistic variations. Second, to achieve fine-grained cross-modal alignment, we develop M2DPO, a retrieval-augmented multi-modal DPO variant that enforces textual preference inversion triggered by visual context switches. Third, we locate correct yet highly uncertain predicted entities and apply counterfactual modifications to construct targeted preference data for latent risk mitigation, thereby further enhancing the model reliability. Extensive experiments on two public chest X-ray datasets (MIMIC-CXR and IU X-Ray) and an in-house endoscopy dataset demonstrate that DPO-Clin significantly improves the SFT baselines on clinical-aware metrics. Furthermore, it achieves superior performance over existing DPO-based MRG methods, exhibiting robust generalizability across distinct baseline architectures and diverse medical imaging modalities.

Figures

Figures reproduced from arXiv: 2608.08494 by the authors.

Figure 1
Figure 1. Comparison between existing DPO-based MRG methods and our DPO-Clin. We utilize distinct colors to indicate descriptions corresponding to different clinical findings. (a) Existing methods construct report preference pairs directly from raw GT and predicted reports, inadvertently entangling critical clinical findings with clin￾ically irrelevant linguistic variations. (b) Guided by the proposed Entity-level Clini￾cal D… view at source ↗
Figure 2
Figure 2. Overview of the proposed DPO-Clin. DPO-Clin is a DPO-based post-training framework designed to optimize the SFT MRG baseline. It leverages an ECD module to assist LLM in generating linguistically-aligned preferred reports y GT, and applies un￾certain entity recognition to obtain counterfactual reports y unc. By incorporating a two￾stage image retrieval strategy, it constructs multi-modal preference data for M2DPO, w… view at source ↗
Figure 3
Figure 3. Visualization of cross-modal attention maps. The attention map represents the average attention weights of all tokens within the un￾derlined sentence [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 23 canonical work pages

  1. [47]

    arXiv preprint arXiv:2406.16845 (2024)

    Zhao, W., Wu, C., Zhang, X., Zhang, Y., Wang, Y., Xie, W.: Ratescore: A metric for radiology report generation. arXiv preprint arXiv:2406.16845 (2024)

    arXiv 2024

  2. [1]

    arXiv preprint arXiv:2303.08774 (2023)

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al.: Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  3. [2]

    arXiv preprint arXiv:2406.06496 (2024)

    Banerjee, O., Zhou, H.Y., Adithan, S., Kwak, S., Wu, K., Rajpurkar, P.: Direct preference optimization for suppressing hallucinated prior exams in radiology re- port generation. arXiv preprint arXiv:2406.06496 (2024)

    arXiv 2024

  4. [3]

    In: Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summa- rization

    Banerjee, S., Lavie, A.: Meteor: An automatic metric for mt evaluation with im- proved correlation with human judgments. In: Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summa- rization. pp. 65–72 (2005)

    2005

  5. [4]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Bu, S., Li, T., Yang, Y., Dai, Z.: Instance-level expert knowledge and aggregate discriminative attention for radiology report generation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 14194– 14204 (2024)

    work page 2024

  6. [5]

    In: Proceedings of the 2020 Conference on Empiri- cal Methods in Natural Language Processing (EMNLP)

    Chen, Z., Song, Y., Chang, T.H., Wan, X.: Generating radiology reports via memory-driven transformer. In: Proceedings of the 2020 Conference on Empiri- cal Methods in Natural Language Processing (EMNLP). pp. 1439–1449 (2020)

    work page 2020

  7. [6]

    IEEE Transactions on Medical Imaging (2025)

    Chen, Z., Bie, Y., Jin, H., Chen, H.: Large language model with region-guided referring and grounding for ct report generation. IEEE Transactions on Medical Imaging (2025)

    work page 2025

  8. [7]

    Journal of the American Medical Informatics Association23(2), 304–310 (2015)

    Demner-Fushman, D., Kohli, M.D., Rosenman, M.B., Shooshan, S.E., Rodriguez, L., Antani, S., Thoma, G.R., McDonald, C.J.: Preparing a collection of radiol- ogy examinations for distribution and retrieval. Journal of the American Medical Informatics Association23(2), 304–310 (2015)

    2015

Show all 49 references
  1. [8]

    arXiv preprint arXiv:2409.00250 (2024)

    Fan, Y., Yang, Z., Liu, R., Li, M., Chang, X.: Medical report generation is a multi- label classification problem. arXiv preprint arXiv:2409.00250 (2024)

    2024 arXiv

  2. [9]

    Hein, D., Chen, Z., Ostmeier, S., Xu, J., Varma, M., Reis, E.P., Michalson, A.E., Bluethgen, C., Shin, H.J., Langlotz, C., et al.: Preference fine-tuning for factuality in chest x-ray interpretation models without human feedback (2024)

    2024

  3. [10]

    Advances in Neural Information Processing Systems37, 107249–107269 (2024)

    Hong, I., Li, Z., Bukharin, A., Li, Y., Jiang, H., Yang, T., Zhao, T.: Adaptive preference scaling for reinforcement learning with human feedback. Advances in Neural Information Processing Systems37, 107249–107269 (2024)

    2024

  4. [11]

    arXiv preprint arXiv:2505.14318 (2025)

    Hou, W., Cheng, Y., Xu, K., Li, H., Hu, Y., Li, W., Liu, J.: Radar: Enhancing ra- diology report generation with supplementary knowledge injection. arXiv preprint arXiv:2505.14318 (2025)

    2025 arXiv

  5. [12]

    Iclr1(2), 3 (2022) 16 Q

    Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W., et al.: Lora: Low-rank adaptation of large language models. Iclr1(2), 3 (2022) 16 Q. Hu et al

    2022

  6. [13]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Hu, Q., Wang, Q., Chen, J., Ji, X., Liu, M., Li, Q., Wang, Z.: Holistic white- light polyp classification via alignment-free dense distillation of auxiliary optical chromoendoscopy. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp....

    2025

  7. [14]

    arXiv preprint arXiv:2601.09209 (2026)

    Hu, Q., Wang, Q., Guo, Y., Li, Q., Wang, Z.: Pairing-free group-level knowledge distillation for robust gastrointestinal lesion classification in white-light endoscopy. arXiv preprint arXiv:2601.09209 (2026)

    2026

  8. [15]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Hu, Q., Yi, Z., Zhou, Y., Peng, F., Liu, M., Li, Q., Wang, Z.: Sali: Short-term align- ment and long-term interaction network for colonoscopy video polyp segmentation. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 531–541. Spri...

    2024

  9. [16]

    arXiv preprint arXiv:2106.14463 (2021)

    Jain, S., Agrawal, A., Saporta, A., Truong, S.Q., Duong, D.N., Bui, T., Chambon, P., Zhang, Y., Lungren, M.P., Ng, A.Y., et al.: Radgraph: Extracting clinical enti- ties and relations from radiology reports. arXiv preprint arXiv:2106.14463 (2021)

    2021 arXiv

  10. [17]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Jin, H., Che, H., Lin, Y., Chen, H.: Promptmrg: Diagnosis-driven prompts for medical report generation. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 2607–2615 (2024)

    2024

  11. [18]

    In: Proceedings of the 56th annual meeting of the association for computational linguistics (volume 1: long papers)

    Jing, B., Xie, P., Xing, E.: On the automatic generation of medical imaging reports. In: Proceedings of the 56th annual meeting of the association for computational linguistics (volume 1: long papers). pp. 2577–2586 (2018)

    2018

  12. [19]

    Scientific data6(1), 317 (2019)

    Johnson, A.E., Pollard, T.J., Berkowitz, S.J., Greenbaum, N.R., Lungren, M.P., Deng, C.y., Mark, R.G., Horng, S.: Mimic-cxr, a de-identified publicly available database of chest radiographs with free-text reports. Scientific data6(1), 317 (2019)

    2019

  13. [20]

    Gastroenterology (2025)

    Jong, M.R., Boers, T.G., Fockens, K.N., Jukema, J.B., Kusters, C.H., Jaspers, T.J., van Heslinga, R.v.E., Slooter, F.C., Struyvenberg, M.R., Bisschops, R., et al.: Gastronet-5m: A multicenter dataset for developing foundation models in gastroin- testinal endoscopy. Gastroenter...

    2025

  14. [21]

    In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Li, M., Lin, B., Chen, Z., Lin, H., Liang, X., Chang, X.: Dynamic graph en- hanced contrastive learningfor chestx-ray report generation. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 3334– 3343 (2023)

    2023

  15. [22]

    In: Pro- ceedings of the 33rd ACM International Conference on Multimedia

    Liang, X., Hu, J., Wang, D., Ma, Z., Zhao, L., Li, R., Wan, B., Wang, Q.: Chexpo: Preference optimization for chest x-ray vlms with counterfactual rationale. In: Pro- ceedings of the 33rd ACM International Conference on Multimedia. pp. 2606–2615 (2025)

    2025

  16. [23]

    In: Text sum- marization branches out

    Lin, C.Y.: Rouge: A package for automatic evaluation of summaries. In: Text sum- marization branches out. pp. 74–81 (2004)

    2004

  17. [24]

    In: Findings of the association for com- putational linguistics: ACL-IJCNLP 2021

    Liu, F., Yin, C., Wu, X., Ge, S., Zhang, P., Sun, X.: Contrastive attention for automatic chest x-ray report generation. In: Findings of the association for com- putational linguistics: ACL-IJCNLP 2021. pp. 269–280 (2021)

    2021

  18. [25]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Liu, H., Wei, D., Xu, Z., Wu, X., Zheng, Y., Wang, L.: Rrg-dpo: Direct preference optimization for clinically accurate radiology report generation. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 552–562. Springer (2025)

    2025

  19. [26]

    arXiv preprint arXiv:1711.05101 (2017)

    Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

    2017 arXiv

  20. [27]

    In: Proceedings of the 40th annual meeting of the Association for Computational Linguistics

    Papineni, K., Roukos, S., Ward, T., Zhu, W.J.: Bleu: a method for automatic evaluation of machine translation. In: Proceedings of the 40th annual meeting of the Association for Computational Linguistics. pp. 311–318 (2002) Linguistically-Aligned and Visually-Grounded Preferenc...

    2002

  21. [28]

    Advances in neural information processing systems36, 53728–53741 (2023)

    Rafailov, R., Sharma, A., Mitchell, E., Manning, C.D., Ermon, S., Finn, C.: Direct preference optimization: Your language model is secretly a reward model. Advances in neural information processing systems36, 53728–53741 (2023)

    2023

  22. [29]

    arXiv preprint arXiv:1707.06347 (2017)

    Schulman, J., Wolski, F., Dhariwal, P., Radford, A., Klimov, O.: Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347 (2017)

    2017 arXiv

  23. [30]

    arXiv preprint arXiv:2402.03300 (2024)

    Shao, Z., Wang, P., Zhu, Q., Xu, R., Song, J., Bi, X., Zhang, H., Zhang, M., Li, Y., Wu, Y., et al.: Deepseekmath: Pushing the limits of mathematical reasoning in open language models. arXiv preprint arXiv:2402.03300 (2024)

    2024 arXiv

  24. [31]

    arXiv preprint arXiv:2004.09167 (2020)

    Smit, A., Jain, S., Rajpurkar, P., Pareek, A., Ng, A.Y., Lungren, M.P.: Chexbert: combining automatic labelers and expert annotations for accurate radiology report labeling using bert. arXiv preprint arXiv:2004.09167 (2020)

    2020 arXiv

  25. [32]

    In: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing

    Sun, G., Qin, C., Fu, H., Wang, L., Tao, Z.: Self-training large language and vision assistant for medical question answering. In: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. pp. 20052–20060 (2024)

    2024

  26. [33]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Tanida,T.,Müller,P.,Kaissis,G.,Rueckert,D.:Interactiveandexplainableregion- guided radiology report generation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 7433–7442 (2023)

    2023

  27. [34]

    In: Findings of the Association for Computational Linguistics: NAACL 2025

    Wang, X., Chen, J., Wang, Z., Zhou, Y., Zhou, Y., Yao, H., Zhou, T., Goldstein, T., Bhatia, P., Kass-Hout, T., et al.: Enhancing visual-language modality alignment in large vision language models via self-improvement. In: Findings of the Association for Computational Linguisti...

    2025

  28. [35]

    arXiv preprint arXiv:2512.02710 (2025)

    Wang, Y., Gao, S., Liu, J., Jiang, S., Xia, H., Zhang, X., Kang, Z., Wang, Y., Liu, Z.: Beyond n-grams: A hierarchical reward learning framework for clinically-aware medical report generation. arXiv preprint arXiv:2512.02710 (2025)

    2025

  29. [36]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Wang, Z., Liu, L., Wang, L., Zhou, L.: Metransformer: Radiology report genera- tion by transformer with multiple learnable expert tokens. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 11558– 11567 (2023)

    2023

  30. [37]

    Meta-Radiology1(3), 100033 (2023)

    Wang, Z., Liu, L., Wang, L., Zhou, L.: R2gengpt: Radiology report generation with frozen llms. Meta-Radiology1(3), 100033 (2023)

    2023

  31. [38]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Wu, C., Zhang, X., Zhang, Y., Wang, Y., Xie, W.: Medklip: Medical knowledge enhanced language-image pre-training for x-ray diagnosis. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 21372–21383 (2023)

    2023

  32. [39]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Yang, K., Tao, J., Lyu, J., Ge, C., Chen, J., Shen, W., Zhu, X., Li, X.: Using human feedback to fine-tune diffusion models without any reward model. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 8941–8951 (2024)

    2024

  33. [40]

    Medical image analysis80, 102510 (2022)

    Yang, S., Wu, X., Ge, S., Zhou, S.K., Xiao, L.: Knowledge matters: Chest radiology report generation with general and specific knowledge. Medical image analysis80, 102510 (2022)

    2022

  34. [41]

    IEEE Transactions on Medical Imaging (2025)

    Yang, Y., You, X., Zhang, K., Fu, Z., Wang, X., Ding, J., Sun, J., Yu, Z., Huang, Q., Han, W., et al.: Spatio-temporal and retrieval-augmented modelling for chest x-ray report generation. IEEE Transactions on Medical Imaging (2025)

    2025

  35. [42]

    Patterns4(9) (2023)

    Yu, F., Endo, M., Krishnan, R., Pan, I., Tsai, A., Reis, E.P., Fonseca, E.K.U.N., Lee, H.M.H., Abad, Z.S.H., Ng, A.Y., et al.: Evaluating progress in automatic chest x-ray radiology report generation. Patterns4(9) (2023)

    2023

  36. [43]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Yu, T., Yao, Y., Zhang, H., He, T., Han, Y., Cui, G., Hu, J., Liu, Z., Zheng, H.T., Sun, M., et al.: Rlhf-v: Towards trustworthy mllms via behavior alignment from fine-grained correctional human feedback. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Patter...

    2024

  37. [44]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

    Yu, T., Zhang, H., Li, Q., Xu, Q., Yao, Y., Chen, D., Lu, X., Cui, G., Dang, Y., He, T., et al.: Rlaif-v: Open-source ai feedback leads to super gpt-4v trustworthiness. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 19985–19995 (2025)

    2025

  38. [45]

    Advances in Neural Information Processing Systems37, 26171–26194 (2024)

    Zhang, M., Wu, W., Lu, Y., Song, Y., Rong, K., Yao, H., Zhao, J., Liu, F., Feng, H., Wang, J., et al.: Automated multi-level preference for mllms. Advances in Neural Information Processing Systems37, 26171–26194 (2024)

    2024

  39. [46]

    In: Proceedings of the AAAI conference on artificial intelligence

    Zhang, Y., Wang, X., Xu, Z., Yu, Q., Yuille, A., Xu, D.: When radiology report generation meets knowledge graph. In: Proceedings of the AAAI conference on artificial intelligence. vol. 34, pp. 12910–12917 (2020)

    2020

  40. [48]

    arXiv preprint arXiv:2403.06728 (2024)

    Zhou, Z., Shi, M., Wei, M., Alabi, O., Yue, Z., Vercauteren, T.: Large model driven radiology report generation with clinical quality reinforcement learning. arXiv preprint arXiv:2403.06728 (2024)

    2024 arXiv

  41. [49]

    In: Forty-second International Conference on Machine Learning (2025)

    Zhu, K., Xia, P., Li, Y., Zhu, H., Wang, S., Yao, H.: Mmedpo: Aligning medical vision-language models with clinical-aware multimodal preference optimization. In: Forty-second International Conference on Machine Learning (2025)

    2025

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.