Pith. sign in

REVIEW 5 major objections 8 minor 44 references

Radiology Report Generation via Multi-objective Preference Optimization

T0 review · 5 major / 8 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 A single radiology-report model can be conditioned by a preference vector to trade off fluency against clinical accuracy, and the paper shows monotone shifts on two public chest X-ray datasets.

desk verdict A new preference-conditioned radiology report generator with a real idea, but the SOTA table appears to cherry-pick across preference vectors and the 'optimal over the entire preference space' claim outruns the evidence. read the letter →

arxiv 2412.08901 v2 pith:CMJPMYNL submitted 2024-12-12 cs.LG cs.AI

Ting Xiao , Lei Shi , Peng Liu , Zhe Wang , Chenjia Bai This is my paper · ORCID
classification cs.LGcs.AI
keywords radiologyreportgenerationmulti-objectivereinforcementlearningpreferencealignmentvectorclinicalefficacychestX-rayIU-XrayMIMIC-CXR
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 tries to establish that radiology report generation can be aligned to heterogeneous, multidimensional radiologist preferences within one model. It proposes MPO: encode preferences as a vector on a simplex, condition the image-to-report transformer on that vector via an attention-fusion module, and train with multi-objective reinforcement learning over randomly sampled preference vectors. The claimed result is an optimal policy over the entire preference space, so at inference the same model can be pointed at any preference, such as fluency versus clinical accuracy, simply by feeding the vector. This matters because different radiologists weight report qualities differently, and prior RL-based methods optimize one fixed reward and must be retrained per preference.

What carries the argument

The argument rests on two coupled mechanisms. The preference vector fusion (PVF) network expands a preference vector $p \in \Delta^{m-1}$ to match encoded visual features and fuses them with multi-head attention, $H_j = \text{Softmax}(P E^\top / \sqrt{d}) E$, followed by a residual connection $U = E + \alpha H$, so that the preference vector conditions the decoder. The multi-objective optimization (MOO) module defines a reward $R = \sum_i p_i r_i(Y)$ as the linear combination of per-metric rewards, and trains by REINFORCE with a self-critical baseline $b$. Training samples preference vectors uniformly across the simplex, which the paper argues yields an optimal policy over the entire preference space.

What would settle it

A reader could fix a preference vector that heavily weights CheXbert-based F1, generate reports on a held-out set, and ask radiologists to rank them against reports generated with a vector heavily weighting BLEU-1; if clinician rankings do not track the weighted clinical dimension, or if a single fixed-reward baseline matches MPO at every point of the trade-off curve, the 'optimal policy over the entire preference space' claim is falsified.

Watch

Extended reading notes

Core claim

The paper claims that a single RRG model can learn a policy over the full preference simplex, so that a low-dimensional preference vector $p$ conditions generation and controls trade-offs among metrics such as BLEU, ROUGE-L, and CheXbert-based F1. The central demonstration is that moving weight between BLEU-1 and F1 on MIMIC-CXR shifts measured clinical F1 from 0.316 to 0.353 while BLEU-1 drops from 0.416 to 0.392, and similar monotone trade-offs appear on IU-Xray. The authors further claim state-of-the-art NLG scores on both datasets and best precision among compared clinical metrics.

Load-bearing premise

The central assumption is that radiologists' preferences can be faithfully represented as weights over the chosen automatic metrics, so that optimizing a linear combination of these rewards aligns generation with any human preference vector.

Editorial extensions

If this is right

  • The same checkpoint can generate reports tuned to different preference trade-offs at inference, because the preference vector is an input condition rather than a retraining target.
  • Raising the preference weight on clinical measures, such as CheXbert-based F1, increases clinical precision, recall, and F1 at the cost of some NLG overlap scores; raising the NLG weight has the opposite effect.
  • A three-dimensional preference vector, including BLEU-1, BLEU-4, and ROUGE-L or F1 and RadCliQ, produces the best score on whichever metric receives weight one, while equal weights act as a balanced average.
  • The design extends to arbitrary automatic metrics as reward dimensions, since the reward is a linear weighting of per-metric improvements.
  • With an equal-weight preference vector, the model behaves as a balanced generator rather than one specialized to any single metric.

Reading between the lines

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

  • Editorial extension: if radiologist preferences are genuinely heterogeneous, this conditional-control scheme could be deployed per reader by storing a small preference vector rather than training per-reader models.
  • Editorial extension: the same preference-conditioning mechanism could be applied to other image-to-text clinical tasks, such as pathology or discharge summaries, by substituting the reward dimensions.
  • Editorial extension: the paper implicitly assumes a linear mapping between metric weights and human value; a direct test of whether radiologists' rankings track the vector-weighted metric remains an open question.
  • Editorial extension: the method's 'optimal policy over the entire preference space' claim could be tested by comparing against a portfolio of fixed-reward RL models at many preference points; if the fixed-reward models dominate the conditional model, the claim would be weakened.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 8 minor

Summary. The paper proposes Multi-objective Preference Optimization (MPO) for radiology report generation. MPO first trains an R2Gen-style encoder-decoder with maximum likelihood, using a Preference Vector Fusion (PVF) module to inject an m-dimensional preference vector into the visual features via attention and residual connections (Eqs. 4-5). It then applies multi-objective reinforcement learning (Eqs. 9-14) to optimize a preference-weighted sum of metric rewards: BLEU-1/ROUGE-L on IU-Xray and BLEU-1/CheXbert F1 on MIMIC-CXR, with preference vectors sampled during training. At inference, the model is conditioned on a user-specified preference vector without further fine-tuning. Experiments report preference-control tables, comparisons with prior RRG methods, ablations, a hyperparameter analysis, and supplementary three-dimensional-preference and RadCliQ experiments.

Significance. If the 'optimal policy over the entire preference space' claim were substantiated, the contribution would be valuable: a single RRG model whose output can be steered along multiple quality dimensions at inference time, avoiding retraining per preference. The paper's positive aspects include the ablation of PVF against simpler fusion methods, the alpha sensitivity analysis, the supplementary three-dimensional preference and RadCliQ experiments, and the efficiency comparison. However, the core evidence currently conflates optimizing automatic metrics with satisfying radiologist preferences, and all tested preference vectors lie on the training grid. With held-out preference vectors, statistical error bars, and at least a small human-preference evaluation, the contribution would be solid.

major comments (5)
  1. [Effective of Preference Guidance, Tables 1-2; Implementation Details; Conclusion] Training samples the preference vector at interval 0.1 and test uses interval 0.2, so every reported test vector (1,0), (0.8,0.2), ..., (0,1) is an in-distribution training point. The observed monotone trends therefore do not support the Conclusion's claim of an 'optimal policy over the entire preference space', which would require generalization to unobserved vectors. Please evaluate held-out preference vectors (e.g., 0.35/0.65 and 0.55/0.45), report interpolation error, and, if possible, compare against per-preference specialist models or a Pareto-frontier analysis.
  2. [Eqs. (9)-(11); Tables 1-4] The RL rewards are BLEU-1 and ROUGE-L on IU-Xray and BLEU-1 and CheXbert F1 on MIMIC-CXR, and the same metric families are used as the evaluation metrics in the main and ablation tables. This makes the preference-control results a check that the model optimizes its training objectives, not that it aligns with radiologists' preferences. The supplementary RadCliQ experiment is a step in the right direction, but RadCliQ is still a learned automatic metric and is not part of the main comparison. Add a human evaluation with radiologists or clinicians rating reports generated under different preference vectors, or at least report a held-out metric not used as a reward.
  3. [Tables 3 and 4] The main comparison tables do not state the preference vector used for MPO. The MIMIC-CXR NLG row in Table 3 matches p=(1,0) of Table 2, while the CE row in Table 4 matches p=(0,1) of Table 2; on IU-Xray the Table 3 row is close to but not identical to the p=(1,0) row in Table 1 (ROUGE-L 0.415 vs. 0.413). Please disclose the exact preference vector(s) used for each reported number, and avoid mixing different vectors across tables in a way that makes the state-of-the-art claim difficult to interpret.
  4. [Tables 1 and 2] None of the preference-control results include error bars or significance tests. The IU-Xray ROUGE-L movement is only 0.413 to 0.415 (0.002 absolute), which is likely within seed-level noise. Report means and standard deviations over multiple random seeds or bootstrap confidence intervals before claiming that the model 'customizes' its behavior.
  5. [Eq. (8)] The per-step reward r_i^t = r_i(Y_1:t) - r_i(Y_1:t-1) is not well-defined for n-gram metrics and clinical labelers when the prefix is shorter than the n-gram order or is an incomplete sentence (e.g., BLEU-4 for t<4, CheXbert on partial text). Since the total telescopes to the sequence reward sum_i p_i r_i(Y), the stepwise decomposition is mathematically redundant; either remove it and use the sequence-level reward directly, or justify why the incremental reward is preferred.
minor comments (8)
  1. [Evaluation Metrics] 'METOR' should be 'METEOR'.
  2. [Implementation Details] 'The sampling interval of the preference vector is 0.1' conflicts with the statement in Methods and Abstract that preference vectors are 'randomly sampled' from the preference space; specify whether training uses a finite grid at resolution 0.1 or continuous i.i.d. sampling.
  3. [Table 4 caption] The caption says 'clinical metrics and diversity scores' but no diversity score is reported; remove the phrase or add the diversity measure.
  4. [Tables 1 and 3] On IU-Xray, MPO's ROUGE-L is 0.415 in Table 3 but 0.413 in Table 1 for the same p=(1,0) configuration; reconcile the discrepancy.
  5. [Comparison with Previous Works] The sentence 'The average value of CE is improved by about 9.4%' is ambiguous between relative improvement and absolute percentage-point improvement; state the calculation explicitly.
  6. [Figure 1] Figure 1 contains a handwritten annotation ('2024.8.16肖婷-1') and garbled placeholder characters; provide a clean figure.
  7. [Related Works] Several preference-alignment citations are arXiv preprints (Yu et al. 2024; Yuan et al. 2024; Shan et al. 2024); clarify how MPO differs from these methods in the RRG context.
  8. [Figure 2] The qualitative preference analysis is based on a single example; consider adding quantitative factual-consistency scores (e.g., CheXbert recall for the highlighted findings) to support the claim that R1 is clinically more faithful.

Circularity Check

1 steps flagged · score 6.0 of 10

Preference-control results re-test the training reward on training-grid preference vectors, making the 'entire preference space' claim partly circular; SOTA benchmarks remain independent.

  1. fitted input called prediction [Methods: Eq. (9)-(11); Training; Experiments: 'Effective of Preference Guidance', Tables 1-2; Implementation Details]
    "During training, the sampling interval of the preference vector on both datasets is 0.1 ... Tables 1 and 2 show the test results with a preference vector sampling interval of 0.2 ... For the IU-Xray dataset, the two rewards in the multi-objective optimization are calculated from the BLEU-1 and ROUGE-L metrics ... as the weight of BLEU-1 decreases from 1 to 0 ... the B1 score decreases from 0.548 to 0.531 ... while the ROUGE-L score increases from 0.413 to 0.415."

    The preference-control 'predictions' are the same quantities used as the training reward. Eq. (9)-(10) define the objective as R = sum_i p_i r_i(Y), with r_i being BLEU-1/ROUGE-L on IU-Xray and BLEU-1/F1 on MIMIC-CXR, and Tables 1-2 report exactly those r_i as evidence of preference alignment. Since the test preference vectors (0.2 grid) are a subset of the 0.1 training grid, the observed monotone changes are entailed by successful optimization of the training reward, not independent evidence for 'an optimal policy on the entire preference space.' No held-out preference vector, human preference rating, or Pareto-frontier analysis is provided.

full rationale

The derivation of PVF and MOO is self-contained: Eqs. (1)-(14) are defined in-paper, the RL update is standard REINFORCE with a self-critical baseline, and the SOTA comparison against external baselines on held-out test images is a genuine benchmark, so the architectural and metric-improvement claims are not circular. The references to the authors' own earlier preference-alignment papers (Yu et al. 2024; Shan et al. 2024; Yuan et al. 2024) are related-work positioning, not load-bearing. The circular component is localized to the preference-control demonstration and the continuous 'entire preference space' conclusion: training samples p on a 0.1 grid, testing re-uses that grid (0.2 subset), and the measured outcomes are the same B1/ROUGE/F1 reward terms that RL was trained to maximize. The paper also never validates the assumed equivalence between radiologist preferences and weights over automatic metrics; that is why the preference-alignment claim reduces, in the evidence presented, to verifying the training objective rather than predicting human preference. Score 6 reflects this partial circularity; the underlying single-model conditioning and held-out image generalization retain independent content.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests mainly on the RL scalarization of automatic metrics as human preferences (unvalidated), plus standard RL estimators and a discretized coverage assumption for the preference simplex. No new physical entities are postulated. Alpha is the only hand-tuned scalar that affects the fusion strength.

free parameters (1)
  • alpha (PVF scaling factor) = 3 on IU-Xray, 0.5 on MIMIC-CXR
    Scaling factor in Eq. (5) controlling the contribution of the preference-fusion attention; values selected by hyperparameter search shown in Table 7.
assumptions (4)
  • standard math The REINFORCE policy gradient with a self-critical baseline provides a valid and sufficiently low-variance gradient estimator for the RL objective.
    Used to derive the parameter update in Eq. (12)-(14). Standard results from Williams (1992) and Rennie et al. (2017).
  • domain assumption Radiologist preferences are representable as weights over automatic metrics (BLEU, ROUGE-L, CheXbert clinical F1), and maximizing their linear combination is equivalent to aligning with the corresponding human preference.
    The entire MOO module depends on this scalarization; no human preference data is collected, so the assumption is untested.
  • domain assumption CheXbert-trained labels provide a reliable ground truth for clinical efficacy, making F1 and recall computed from them valid measures of report quality.
    Used for the clinical efficacy reward and evaluation (Experiments, Evaluation Metrics).
  • ad hoc to paper Randomly sampling preference vectors from the simplex during training yields an optimal policy for all points in the continuous preference space.
    The paper states this in the Training section without a coverage or convergence argument; the experiments only evaluate a discrete grid.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Radiology Report Generation via Multi-objective Preference Optimization." pith.science (2026). https://pith.science/paper/CMJPMYNL

@misc{pith2026241208901,
  author       = {Pith},
  title        = {Pith review of: Radiology Report Generation via Multi-objective Preference Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CMJPMYNL}},
  note         = {Machine review of arXiv:2412.08901}
}
read the original abstract

Automatic Radiology Report Generation (RRG) is an important topic for alleviating the substantial workload of radiologists. Existing RRG approaches rely on supervised regression based on different architectures or additional knowledge injection,while the generated report may not align optimally with radiologists' preferences. Especially, since the preferences of radiologists are inherently heterogeneous and multidimensional, e.g., some may prioritize report fluency, while others emphasize clinical accuracy. To address this problem,we propose a new RRG method via Multi-objective Preference Optimization (MPO) to align the pre-trained RRG model with multiple human preferences, which can be formulated by multi-dimensional reward functions and optimized by multi-objective reinforcement learning (RL). Specifically, we use a preference vector to represent the weight of preferences and use it as a condition for the RRG model. Then, a linearly weighed reward is obtained via a dot product between the preference vector and multi-dimensional reward. Next,the RRG model is optimized to align with the preference vector by optimizing such a reward via RL. In the training stage,we randomly sample diverse preference vectors from the preference space and align the model by optimizing the weighted multi-objective rewards, which leads to an optimal policy on the entire preference space. When inference,our model can generate reports aligned with specific preferences without further fine-tuning. Extensive experiments on two public datasets show the proposed method can generate reports that cater to different preferences in a single model and achieve state-of-the-art performance.

Figures

Figures reproduced from arXiv: 2412.08901 by the authors.

Figure 1
Figure 1. The architecture of our MPO, where the blue dashed box represents the encoder-decoder with the PVF network, and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Reports from ground truth and MPO with different [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 1
Figure 1. Reports from ground truth and MPO with different [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 29 canonical work pages

  1. [1]

    Cao, Y.; Cui, L.; Zhang, L.; Yu, F.; Li, Z.; and Xu, Y. 2023. MMTN: multi-modal memory transformer network for image-report consistent medical report generation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 277--285

    work page 2023

  2. [2]

    Chen, Z.; Shen, Y.; Song, Y.; and Wan, X. 2021. Cross-modal Memory Networks for Radiology Report Generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), 5904--5914

    work page 2021

  3. [3]

    Chen, Z.; Song, Y.; Chang, T.-H.; and Wan, X. 2020. Generating Radiology Reports via Memory-driven Transformer. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), 1439--1449

    work page 2020

  4. [4]

    Delbrouck, J.-B.; Chambon, P.; Bluethgen, C.; Tsai, E.; Almusa, O.; and Langlotz, C. 2022. Improving the Factual Correctness of Radiology Report Generation with Semantic Rewards. In Findings of the Association for Computational Linguistics: EMNLP 2022, 4348--4360

    work page 2022

  5. [5]

    D.; Rosenman, M

    Demner-Fushman, D.; Kohli, M. D.; Rosenman, M. B.; Shooshan, S. E.; Rodriguez, L.; Antani, S.; Thoma, G. R.; and McDonald, C. J. 2016. Preparing a collection of radiology examinations for distribution and retrieval. Journal of the American Medical Informatics Association, 23(2): 304--310

    work page 2016

  6. [6]

    Gu, T.; Liu, D.; Li, Z.; and Cai, W. 2024. Complex Organ Mask Guided Radiology Report Generation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 7995--8004

    work page 2024

  7. [7]

    Gu, Y.; Tinn, R.; Cheng, H.; Lucas, M.; Usuyama, N.; Liu, X.; Naumann, T.; Gao, J.; and Poon, H. 2021. Domain-specific language model pretraining for biomedical natural language processing. ACM Transactions on Computing for Healthcare (HEALTH), 3(1): 1--23

    2021

  8. [8]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

    2016

Show all 44 references
  1. [9]

    Huang, Z.; Zhang, X.; and Zhang, S. 2023. Kiut: Knowledge-injected u-transformer for radiology report generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 19809--19818

    2023

  2. [10]

    Jin, H.; Che, H.; Lin, Y.; and Chen, H. 2024. Promptmrg: Diagnosis-driven prompts for medical report generation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 2607--2615

    2024

  3. [11]

    Jing, B.; Xie, P.; and Xing, E. 2018. 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), 2577--2586

    2018

  4. [12]

    E.; Pollard, T

    Johnson, A. E.; Pollard, T. J.; Greenbaum, N. R.; Lungren, M. P.; Deng, C.-y.; Peng, Y.; Lu, Z.; Mark, R. G.; Berkowitz, S. J.; and Horng, S. 2019. MIMIC-CXR-JPG, a large publicly available database of labeled chest radiographs. arXiv preprint arXiv:1901.07042

    2019 arXiv

  5. [13]

    Knowledge is Power

    Kale, K.; Bhattacharyya, P.; Shetty, A.; Gune, M.; Shrivastava, K.; Lawyer, R.; and Biswas, S. 2023. “Knowledge is Power”: Constructing Knowledge Graph of Abdominal Organs and Using Them for Automatic Radiology Report Generation. In Proceedings of the 61st Annual Meeting of th...

    2023

  6. [14]

    Li, Y.; Liang, X.; Hu, Z.; and Xing, E. P. 2018. Hybrid retrieval-generation reinforced agent for medical image report generation. Advances in neural information processing systems, 31

    2018

  7. [15]

    Lin, C.-Y. 2004. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, 74--81

    2004

  8. [16]

    Liu, A.; Guo, Y.; Yong, J.-h.; and Xu, F. 2024 a . Multi-grained Radiology Report Generation with Sentence-level Image-language Contrastive Learning. IEEE Transactions on Medical Imaging

    2024

  9. [17]

    Liu, C.; Tian, Y.; Chen, W.; Song, Y.; and Zhang, Y. 2024 b . Bootstrapping Large Language Models for Radiology Report Generation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 18635--18643

    2024

  10. [18]

    Liu, F.; Wu, X.; Ge, S.; Fan, W.; and Zou, Y. 2021 a . Exploring and distilling posterior and prior knowledge for radiology report generation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 13753--13762

    2021

  11. [19]

    Liu, F.; Yin, C.; Wu, X.; Ge, S.; Zhang, P.; and Sun, X. 2021 b . Contrastive Attention for Automatic Chest X-ray Report Generation. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, 269--280

    2021

  12. [20]

    Miura, Y.; Zhang, Y.; Tsai, E.; Langlotz, C.; and Jurafsky, D. 2021. Improving Factual Completeness and Consistency of Image-to-Text Radiology Report Generation. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistic...

    2021

  13. [21]

    Najdenkoska, I.; Zhen, X.; Worring, M.; and Shao, L. 2021. Variational topic inference for chest x-ray report generation. In Medical Image Computing and Computer Assisted Intervention--MICCAI 2021: 24th International Conference, Strasbourg, France, September 27--October 1, 202...

    2021

  14. [22]

    Papineni, K.; Roukos, S.; Ward, T.; and Zhu, W. 2002. Bleu: a Method for Automatic Evaluation of Machine Translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, July 6-12, 2002, Philadelphia, PA, USA , 311--318. ACL

    2002

  15. [23]

    Qin, H.; and Song, Y. 2022. Reinforced cross-modal alignment for radiology report generation. In Findings of the Association for Computational Linguistics: ACL 2022, 448--458

    2022

  16. [24]

    J.; Marcheret, E.; Mroueh, Y.; Ross, J.; and Goel, V

    Rennie, S. J.; Marcheret, E.; Mroueh, Y.; Ross, J.; and Goel, V. 2017. Self-critical sequence training for image captioning. In Proceedings of the IEEE conference on computer vision and pattern recognition, 7008--7024

    2017

  17. [25]

    Shan, Z.; Fan, C.; Qiu, S.; Shi, J.; and Bai, C. 2024. Forward kl regularized preference optimization for aligning diffusion policies. arXiv preprint arXiv:2409.05622

    2024 arXiv

  18. [26]

    Shen, H.; Pei, M.; Liu, J.; and Tian, Z. 2024. Automatic Radiology Reports Generation via Memory Alignment Network. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 4776--4783

    2024

  19. [27]

    Y.; and Lungren, M

    Smit, A.; Jain, S.; Rajpurkar, P.; Pareek, A.; Ng, A. Y.; and Lungren, M. 2020. Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMN...

    2020

  20. [28]

    Song, X.; Zhang, X.; Ji, J.; Liu, Y.; and Wei, P. 2022. Cross-modal contrastive attention model for medical report generation. In Proceedings of the 29th International Conference on Computational Linguistics, 2388--2397

    2022

  21. [29]

    Tanida, T.; M \"u ller, P.; Kaissis, G.; and Rueckert, D. 2023. Interactive and explainable region-guided radiology report generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7433--7442

    2023

  22. [30]

    N.; Kaiser, .; and Polosukhin, I

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

    2017

  23. [31]

    Wang, F.; Liang, X.; Xu, L.; and Lin, L. 2020. Unifying relational sentence generation and retrieval for medical image report composition. IEEE transactions on cybernetics, 52(6): 5015--5025

    2020

  24. [32]

    Wang, J.; Bhalerao, A.; and He, Y. 2022. Cross-modal prototype driven network for radiology report generation. In European Conference on Computer Vision, 563--579. Springer

    2022

  25. [33]

    Wang, Z.; Liu, L.; Wang, L.; and Zhou, L. 2023. Metransformer: Radiology report generation by transformer with multiple learnable expert tokens. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11558--11567

    2023

  26. [34]

    Wang, Z.; Tang, M.; Wang, L.; Li, X.; and Zhou, L. 2022. A medical semantic-assisted transformer for radiographic report generation. In International Conference on Medical Image Computing and Computer-Assisted Intervention, 655--664. Springer

    2022

  27. [35]

    Wang, Z.; Zhou, L.; Wang, L.; and Li, X. 2021. A self-boosting framework for automated radiographic report generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2433--2442

    2021

  28. [36]

    Williams, R. J. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning, 8: 229--256

    1992

  29. [37]

    Wu, Y.; Huang, I.-C.; and Huang, X. 2023. Token Imbalance Adaptation for Radiology Report Generation. In Conference on Health, Inference, and Learning, 72--85. PMLR

    2023

  30. [38]

    K.; and Xiao, L

    Yang, S.; Wu, X.; Ge, S.; Zheng, Z.; Zhou, S. K.; and Xiao, L. 2023. Radiology report generation with a learned knowledge base and multi-modal alignment. Medical Image Analysis, 86: 102798

    2023

  31. [39]

    K.; and Xiao, L

    Yang, S.; Wu, X.; Ge, S.; Zhou, S. K.; and Xiao, L. 2022. Knowledge matters: Chest radiology report generation with general and specific knowledge. Medical image analysis, 80: 102510

    2022

  32. [40]

    You, D.; Liu, F.; Ge, S.; Xie, X.; Zhang, J.; and Wu, X. 2021. Aligntransformer: Hierarchical alignment of visual regions and disease tags for medical report generation. In Medical Image Computing and Computer Assisted Intervention--MICCAI 2021: 24th International Conference, ...

    2021

  33. [41]

    Yu, X.; Bai, C.; He, H.; Wang, C.; and Li, X. 2024. Regularized Conditional Diffusion Model for Multi-Task Preference Alignment. arXiv preprint arXiv:2404.04920

    2024 arXiv

  34. [42]

    Yuan, X.; Shang, Z.; Wang, Z.; Wang, C.; Shan, Z.; Qi, Z.; Zhu, M.; Bai, C.; and Li, X. 2024. Preference Aligned Diffusion Planner for Quadrupedal Locomotion Control. arXiv preprint arXiv:2410.13586

    2024 arXiv

  35. [43]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  36. [44]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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