Histopathology Image Report Generation by Vision Language Model with Multimodal In-Context Learning

Fu-En Yang; Hsuan-Yu Fan; Shih-Wen Liu; Wei-Ta Chu; Yu-Chiang Frank Wang

REVIEW 3 major objections 3 minor 16 references

Histopathology Image Report Generation by Vision Language Model with Multimodal In-Context Learning

T0 review · 3 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read PathGenIC claims top HistGen report scores by adding retrieved examples, guidelines, and feedback to a vision-language model.

desk verdict Clear writing and a sensible ablation, but the SOTA claim depends on an evaluation that gives PathGenIC context no baseline gets; the measured gains are small and unquantified. read the letter →

arxiv 2506.17645 v1 pith:2XZFJ5DN submitted 2025-06-21 cs.CV

Shih-Wen Liu , Hsuan-Yu Fan , Wei-Ta Chu , Fu-En Yang , Yu-Chiang Frank Wang This is my paper · ORCID

classification cs.CV

keywords histopathologyreportgenerationmultimodalin-contextlearningvision-languagemodelwholeslideimageretrieval-augmentedGPT-4ofeedbackHistGenbenchmarkLoRAfine-tuning

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that histopathology report generation improves substantially if, before writing a report for a whole slide image, the model is shown similar past cases the way a pathologist would consult them. It proposes PathGenIC, which retrieves the nearest-neighbor WSI-report pair from the training set, obtains a disease-category guideline and corrective feedback produced by GPT-4o, and folds all three into the prompt of a fine-tuned vision-language model. On the HistGen benchmark, the full system reports BLEU-1 0.431, BLEU-4 0.196, METEOR 0.197, ROUGE-L 0.357, and factENT 0.462, the strongest published numbers among the compared methods. The relevance is that report writing, not just image classification, is a bottleneck in pathology workflows, and this is a concrete recipe for extracting more from training data at inference time.

What carries the argument

The load-bearing mechanism is multimodal in-context learning with three context signals: nearest-neighbor retrieval, where the test WSI's contextualized tokens are concatenated with the tokens of the most similar training WSI and its ground-truth report; category guidelines, where GPT-4o summarizes report-writing habits from the top-K nearest disease category; and feedback, where GPT-4o critiques base-model drafts against truth and the critiques of the K nearest training cases are fed in. These context cues are injected through the text prompt and token stream of a LoRA-adapted Quilt-LLaVA, whose visual front end is HistGen's DINOv2 ViT-L with learnable query tokens.

What would settle it

Run the Table 1 comparison again but append the same top-1 nearest-neighbor ground-truth report to the prompt of every baseline, including HistGen and the base model, while keeping all other settings fixed; if PathGenIC's BLEU-4 advantage over HistGen shrinks to roughly zero, the state-of-the-art claim is explained by the added retrieval context rather than by the method itself.

Watch

Extended reading notes

Core claim

The central claim is that in-context learning over the training set, rather than a larger architecture or more parameters, is what lifts report generation quality. Starting from Quilt-LLaVA, the authors replace its visual encoder with HistGen's DINOv2 ViT-L, add learnable query tokens and transformer blocks to make fixed-length WSI tokens, and LoRA-tune the VLM. On top of this base, three context signals are appended: the image tokens and report of the most cosine-similar training WSI, a GPT-4o-generated guideline distilled from up to 20 reports in the nearest disease category, and GPT-4o feedback that compares base-generated training reports against ground truth. Each component alone helps; together they produce the reported state-of-the-art numbers, and the paper shows the gain is not a quirk of report length or disease category.

Load-bearing premise

The comparison is only meaningful if giving PathGenIC retrieved example reports, GPT-4o disease guidelines, and GPT-4o feedback at test time does not hand it an advantage that the comparison models were never offered.

Editorial extensions

If this is right

If the reported numbers hold, any report generation system on HistGen can be improved by retrieving similar cases at test time without retraining the visual encoder.
The K=3 retrieval setting gives the best tradeoff, so the method is robust to the number of retrieved neighbors rather than requiring exact tuning.
The per-category analysis suggests the gains are consistent across disease types, so the method is not riding on one easy subset.
A direct corollary of the ablations is that each context signal contributes incrementally, with the full combination giving the largest improvement.
Because the base model already outperforms HistGen, part of the state-of-the-art result comes from the improved backbone, while the in-context components provide the additional gains in Table 3.

Reading between the lines

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

Editorial inference: The reported comparison is not apples-to-apples, because PathGenIC receives GPT-4o-generated text at test time while the baselines receive only the test image; a control that gives every baseline the same retrieved report would determine how much of the gain comes from retrieval versus from GPT-4o's prompt-writing ability.
Editorial inference: The same recipe could transfer to other dense-report domains such as radiology, dermatopathology, or surgical pathology notes, provided a retrieval index and a source of supervisory feedback exist.
Editorial inference: The category guideline prompt asks GPT-4o to expose report-writing biases, which suggests the method could amplify systematic wording or coding biases present in a hospital's historical reports; measuring entity-level calibration rather than only BLEU overlap would test this.
Editorial inference: Because evaluation truncates reports at 100 tokens, the metric gains may not reflect full-report clinical completeness; a clinician-rated study on full-length reports would be the natural next test.

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

Here's my take on this paper.

The useful part is real: the authors take Quilt-LLaVA, swap in HistGen's ViT-L encoder, and show that nearest-neighbor retrieval of WSI-report pairs improves BLEU scores on the HistGen benchmark. That specific combination is not in the cited literature, and the paper is transparent about its prompts and ablations. The ablation in Table 3 is particularly clear: the retrieval component delivers most of the improvement (0.411 to 0.428 BLEU-1), and the guideline and feedback components each add about a thousandth. That's honest and reproducible in spirit.

The soft spots are mostly about the evaluation protocol. The "state-of-the-art" claim rests on a comparison where PathGenIC is the only system that receives retrieved reports, GPT-4o category guidelines, and GPT-4o feedback at test time. No baseline gets an equivalent retrieval prompt or extra context, so the comparison measures a full system against stripped baselines, not the method against comparable methods. The feedback component is the trickiest: it is generated by asking GPT-4o to compare the base model's training outputs to ground-truth reports, and at test time the model sees feedback for retrieved training WSIs. That is not test-set leakage, but it is a supervision-derived hint that no baseline receives, and it makes the "fair comparison" claim hard to swallow. Also, the margins over HistGen are tiny——0.018 BLEU-1 and 0.012 BLEU-4——and there are no error bars or significance tests, so I'm not confident those differences are meaningful. The factENT column is reported only for their own models, which weakens the one metric where the gap looks large. Finally, no code is released, though the appendix does include the exact prompts.

None of this is fatal if the claim is softened. The paper is a coherent engineering study, and the results are likely reproducible. But as written, "state-of-the-art" overstates what is demonstrated. A referee should ask for a baseline that also receives retrieved examples and guidelines, repeated runs with error bars, and a clear statement about the privileged information in the feedback.

Recommendation: this deserves a serious referee. The core idea is sound and the issues are identifiable and fixable. I would send it to review with the expectation of major revision, not rejection.

Referee Report

3 major / 3 minor

Summary. The paper proposes PathGenIC, a framework for histopathology report generation from whole slide images. The method builds on Quilt-LLaVA with HistGen's DINOv2 ViT-L feature extractor and introduces three in-context learning components: nearest-neighbor retrieval of WSI-report pairs, GPT-4o-generated category guidelines, and GPT-4o-generated corrective feedback derived by comparing base-model outputs against ground-truth training reports. The framework is evaluated on the HistGen benchmark, and the paper claims state-of-the-art results, reporting BLEU-1 0.431, BLEU-4 0.196, METEOR 0.197, ROUGE-L 0.357, and factENT 0.462 for the full PathGenIC model, with ablations over components, the number of retrieved neighbors, report length, and disease categories.

Significance. If the reported results are robust, the paper makes a useful empirical contribution: it demonstrates that retrieval-based in-context learning, category-level guidelines, and feedback derived from training-set comparisons can improve vision-language report generation on whole slide images. The clinical motivation is well grounded, the method description is clear, and the appendices provide the exact prompts used. The ablations and the analysis across disease categories are also valuable. However, the core claim of state-of-the-art performance is weakened by an asymmetric evaluation protocol and the absence of statistical significance measures. PathGenIC is the only method receiving retrieved reports, GPT-4o-generated guidelines, and feedback derived from ground-truth comparisons, so the reported gains may partly reflect additional test-time information rather than the method itself. The improvement margins over HistGen are small, and no error bars or significance tests are reported. These issues prevent the paper from currently supporting its strongest claim, but they are addressable with additional controls and analysis.

major comments (3)

[§3.2, Eq. (5), Table 1] The comparison in Table 1 is not apples-to-apples. PathGenIC is the only system that receives (i) retrieved WSI tokens and their associated reports, (ii) GPT-4o-generated category guidelines, and (iii) GPT-4o feedback produced by comparing the base model's output with the ground-truth training reports (Eq. (5)). None of the baselines is given an equivalent retrieval/reference/feedback prompt, so the reported improvements (BLEU-1 0.431 vs 0.413, BLEU-4 0.196 vs 0.184) may come from the extra test-time information rather than from the ICL mechanism. In particular, feedback B_j is derived from ground-truth reports of the retrieved training WSIs and can state which entities the generated report lacks; exposing this to the VLM at test time is a form of privileged information. I do not claim mathematical circularity, since no test ground truth is used, but the 'state-of-the-art' claim is only valid under this asymmetric protocol. Please add controls where the same retrieved reports, guidelines, and feedback are provided to the baselines (e.g., HistGen with identical ICL context), and include a random-retrieval condition to quantify the contribution of retrieval quality.
[§4.1, §4.2, Table 1 and Table 3] No measure of variability is reported. All numbers appear to be from single runs; the margins over HistGen are small (BLEU-1 +0.018, BLEU-4 +0.012, METEOR +0.015, ROUGE-L +0.013), and Table 3 shows that adding the guideline and feedback components to nearest-neighbor retrieval changes BLEU-1 by 0.003 and ROUGE-L by 0.003. Without multiple seeds, bootstrap confidence intervals, or significance tests, the claim of 'significant improvements across BLEU, METEOR, and ROUGE-L' in the abstract is unsupported. Additionally, factENT is given only for the two proposed models, with the text noting that values for other methods are 'not available' in §4.2, so the factual-completeness advantage cannot be verified. Please provide standard deviations or significance tests and either obtain factENT for the baselines or remove the implied comparative claim.
[§4.2, Figure 3 and Table 1] The paper states that the results in Table 1 are obtained 'based on only the first 100 tokens of the generated reports.' It is unclear whether this truncation was applied to the reported baseline numbers taken from HistGen or only to the proposed models. If the baselines were evaluated on full generated reports, the comparison is biased because BLEU, METEOR, and ROUGE-L are sensitive to report length. Please clarify explicitly whether the same truncation was applied to every method, and ideally report results on full generated reports as well.

minor comments (3)

[§3.2 and Appendix A] The category guideline procedure is described as using the K most similar WSIs to determine the majority category, but the prompt in Appendix A lists 'Report 1 ... Report 20'. The relationship between K and the number of reports used to generate the guideline is not described; please clarify how the 20 reports are selected.
[Figure 4] The bar plot in Figure 4 shows BLEU scores across 32 disease categories, but no error bars or sample sizes per category are provided, which makes it hard to determine which cross-category differences are meaningful.
[References and Reproducibility] The motivation cites 'Doe and Smith, 2023' with a webpage reference; a more standard peer-reviewed source on pathology peer review would strengthen the introduction. In addition, the paper does not report the GPT-4o sampling parameters or the computational cost of the feedback and guideline generation, which would aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: PathGenIC's test-time context comes from training data only, so the SOTA claim may be an evaluation-fairness concern but not a self-referential derivation.

full rationale

The paper's derivation chain is: (1) take Quilt-LLaVA with HistGen's frozen ViT-L; (2) fine-tune transformer/projector/LoRA on HistGen training reports with cross-entropy; (3) at test time retrieve the nearest training WSI-report pair by cosine similarity to H_test (Eq. 3), gather GPT-4o category guidelines from training reports (Eq. 4), and retrieve GPT-4o feedback generated from base-model outputs and ground-truth reports on training WSIs (Eq. 5). None of these inputs uses the test ground-truth report: retrieval is H_test vs H_train, guidelines are from category C of nearest training WSIs, and feedback B_i is precomputed on training data. Thus the generated test report is not definitionally forced by the target; the prediction is not a fitted input renamed as a prediction. The paper includes no uniqueness theorem and no load-bearing self-citation; all cited methods (HistGen, Quilt-LLaVA, R2Gen, etc.) are external. The only flagged weaknesses are in the evaluation protocol: Table 1 reports factENT only for the proposed methods ('the values of other methods are not available'), and baselines are not given the retrieved prompts; these affect whether the SOTA claim is an apples-to-apples comparison, but they do not make the derivation circular. Section 4.3's admission that key entities are still missed is a limitation statement, not a circular step. Verdict: no significant circularity.

Assumptions & free parameters 4 free parameters · 3 assumptions · 1 invented entities

The free parameters are standard empirical choices for a deep learning paper (query token count, retrieval K, training hyperparameters, report length cutoff). The important assumptions are about retrieval informativeness and GPT-4o output reliability, both untested. The paper introduces no new entities.

free parameters (4)

Number of query tokens m = not stated
The number of learnable query tokens used to compress each WSI is a design choice that is not reported in the paper, although the scheme depends on it.
K for nearest-neighbor retrieval = 3
The number of retrieved neighboring cases is tuned on the benchmark: Table 2 shows K=1, K=3, and K=5, with K=3 reported as best. This is a fitted hyperparameter, though the differences are small and no significance test is given.
Training hyperparameters = batch size 8, 20 epochs, lr 1e-4, cosine schedule
The learning rate, schedule, epochs, and batch size are chosen empirically to 'ensure performance convergence', per Section 4.1, without ablations.
Report length cutoff = first 100 tokens
The headline scores in Table 1 are computed on the first 100 tokens only, a per-paper evaluation choice that affects the reported magnitudes.

assumptions (3)

domain assumption The retrieved nearest-neighbor reports are genuinely informative for generating the target report.
Section 3.2 assumes that image-token similarity retrieves reports whose content is useful for the target WSI. No analysis of retrieval quality is provided, only final metric changes.
domain assumption GPT-4o generated guidelines and feedback are accurate and transferable to the target WSI.
The method relies on GPT-4o outputs at train time to produce category guidelines and feedback, with no verification of their factual correctness. The prompts even instruct the model to 'be honest and harsh', introducing variability.
domain assumption Lexical overlap metrics (BLEU, METEOR, ROUGE-L) correlate with report quality on this task.
Section 4.1 concedes these metrics may not reflect domain entities or inferential consistency, yet the main state-of-the-art claim rests on them. The factENT metric is only reported for the authors' own systems.

invented entities (1)

None
purpose: None
The paper introduces no new particles, mediators, forces, or geometric structures. The only new artifact is the PathGenIC pipeline itself.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Histopathology Image Report Generation by Vision Language Model with Multimodal In-Context Learning." pith.science (2026). https://pith.science/paper/2XZFJ5DN

@misc{pith2026250617645,
  author       = {Pith},
  title        = {Pith review of: Histopathology Image Report Generation by Vision Language Model with Multimodal In-Context Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2XZFJ5DN}},
  note         = {Machine review of arXiv:2506.17645}
}

read the original abstract

Automating medical report generation from histopathology images is a critical challenge requiring effective visual representations and domain-specific knowledge. Inspired by the common practices of human experts, we propose an in-context learning framework called PathGenIC that integrates context derived from the training set with a multimodal in-context learning (ICL) mechanism. Our method dynamically retrieves semantically similar whole slide image (WSI)-report pairs and incorporates adaptive feedback to enhance contextual relevance and generation quality. Evaluated on the HistGen benchmark, the framework achieves state-of-the-art results, with significant improvements across BLEU, METEOR, and ROUGE-L metrics, and demonstrates robustness across diverse report lengths and disease categories. By maximizing training data utility and bridging vision and language with ICL, our work offers a solution for AI-driven histopathology reporting, setting a strong foundation for future advancements in multimodal clinical applications.

Figures

Figures reproduced from arXiv: 2506.17645 by the authors.

**Figure 1.** Overview of the proposed framework. WSI patch features Fpatch and the learnable tokens Q are jointly processed by a transformer to get holistic WSI features H. The processed H with the text prompts Ptext are then fed to a VLM to generate a report Ygen. Our main contribution is enriching prompts with extra context. The components with flame symbols mean that we need to train or fine-tune the parameters, and the ones … view at source ↗

**Figure 2.** Illustrations of three different clues for in-context learning. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Performance across varying sequence lengths (100 to 500 tokens). [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: BLEU scores across the 32 disease categories. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: A sample result of the generated report and its corresponding ground truth. The generated result is relatively better [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

**Figure 6.** Figure 6: A sample result of the generated report and its corresponding ground truth. The generated result is relatively worse. formance in [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 15 canonical work pages

[1]

Bottom-up and top-down attention for image captioning and visual question answering

Peter Anderson, Xiaodong He, Chris Buehler, Damien Teney, Mark Johnson, Stephen Gould, and Lei Zhang. Bottom-up and top-down attention for image captioning and visual question answering. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6077--6086, 2017

work page 2017
[2]

Improving diagnostic accuracy through feedback: The diagnosis learning cycle

Carolina Fernandez Branson, Michelle Williams, Teresa M Chan, Mark L Graber, Kathleen P Lane, Skip Grieser, Zach Landis-Lewis, James Cooke, Divvy K Upadhyay, Shawn Mondoux, Hardeep Singh, Laura Zwaan, Charles Friedman, and Andrew P J Olson. Improving diagnostic accuracy through feedback: The diagnosis learning cycle. BMJ Quality & Safety, 30 0 (12): 0 100...

work page 2021
[3]

Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin...

work page 1901
[4]

Wsicaption: Multiple instance generation of pathology reports for gigapixel whole-slide images

Pingyi Chen, Honglin Li, Chenglu Zhu, Sunyi Zheng, Zhongyi Shui, and Lin Yang. Wsicaption: Multiple instance generation of pathology reports for gigapixel whole-slide images. In Proceedings of International Conference on Medical Image Computing and Computer Assisted Intervention, pages 546--556, 2024

work page 2024
[5]

Generating radiology reports via memory-driven transformer

Zhihong Chen, Yan Song, Tsung-Hui Chang, and Xiang Wan. Generating radiology reports via memory-driven transformer. In Proceedings of Conference on Empirical Methods in Natural Language Processing, pages 1439--1449, 2020

work page 2020
[6]

Cross-modal memory networks for radiology report generation

Zhihong Chen, Yaling Shen, Yan Song, and Xiang Wan. Cross-modal memory networks for radiology report generation. In Proceedings of Annual Meeting of the Association for Computational Linguistics, pages 5904--5914, 2021

work page 2021
[7]

Meshed-memory transformer for image captioning

Marcella Cornia, Matteo Stefanini, Lorenzo Baraldi, and Rita Cucchiara. Meshed-memory transformer for image captioning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10578--10587, 2020

work page 2020
[8]

Anatomic pathology quality assurance through peer review

John Doe and Alice Smith. Anatomic pathology quality assurance through peer review. PathologyOutlines, 2023. https://www.pathologyoutlines.com/topic/managementlabAPQA.html

work page 2023

Show all 16 references

[9]

Histgen: A local-global encoding framework for pathology report generation

Zhengrui Guo, Jiabo Ma, Yingxue Xu, Yihui Wang, Liansheng Wang, and Hao Chen. Histgen: A local-global encoding framework for pathology report generation. In Proceedings of Medical Image Computing and Computer Assisted Intervention, pages 189--199, 2024

2024
[10]

Lora: Low-rank adaptation of large language models

Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. In Proceedings of International Conference on Learning Representations, 2022

2022
[11]

Retrieval-augmented generation for knowledge-intensive nlp tasks

Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. Retrieval-augmented generation for knowledge-intensive nlp tasks. In Proceedings of Con...

2020
[12]

Llava-med: Training a large language-and-vision assistant for biomedicine in one day

Chunyuan Li, Cliff Wong, Sheng Zhang, Naoto Usuyama, Haotian Liu, Jianwei Yang, Tristan Naumann, Hoifung Poon, and Jianfeng Gao. Llava-med: Training a large language-and-vision assistant for biomedicine in one day. In Proceedings of Conference on Neural Information Processing ...

2023
[13]

Improving factual completeness and consistency of image-to-text radiology report generation

Yasuhide Miura, Yuhao Zhang, Emily Tsai, Curtis Langlotz, and Dan Jurafsky. Improving factual completeness and consistency of image-to-text radiology report generation. In Proceedings of the Conference of the North American Chapter of the Association for Computational Linguist...

2021
[14]

Ikezogwo, Fatemeh Ghezloo, Ranjay Krishna, and Linda Shapiro

Mehmet Saygin Seyfioglu, Wisdom O. Ikezogwo, Fatemeh Ghezloo, Ranjay Krishna, and Linda Shapiro. Quilt-llava: Visual instruction tuning by extracting localized narratives from open-source histopathology videos. arXiv:2312.04746v3, 2025

2025 arXiv
[15]

Gomez, Lukasz Kaiser, and Illia Polosukhin

Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Proceedings of Conference on Neural Information Processing Systems, 2017

2017
[16]

Show and tell: A neural image caption generator

Oriol Vinyals, Alexander Toshev, Samy Bengio, and Dumitru Erhan. Show and tell: A neural image caption generator. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pages 3156--3164, 2015

2015

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Bottom-up and top-down attention for image captioning and visual question answering

Peter Anderson, Xiaodong He, Chris Buehler, Damien Teney, Mark Johnson, Stephen Gould, and Lei Zhang. Bottom-up and top-down attention for image captioning and visual question answering. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6077--6086, 2017

work page 2017

[2] [2]

Improving diagnostic accuracy through feedback: The diagnosis learning cycle

Carolina Fernandez Branson, Michelle Williams, Teresa M Chan, Mark L Graber, Kathleen P Lane, Skip Grieser, Zach Landis-Lewis, James Cooke, Divvy K Upadhyay, Shawn Mondoux, Hardeep Singh, Laura Zwaan, Charles Friedman, and Andrew P J Olson. Improving diagnostic accuracy through feedback: The diagnosis learning cycle. BMJ Quality & Safety, 30 0 (12): 0 100...

work page 2021

[3] [3]

Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin...

work page 1901

[4] [4]

Wsicaption: Multiple instance generation of pathology reports for gigapixel whole-slide images

Pingyi Chen, Honglin Li, Chenglu Zhu, Sunyi Zheng, Zhongyi Shui, and Lin Yang. Wsicaption: Multiple instance generation of pathology reports for gigapixel whole-slide images. In Proceedings of International Conference on Medical Image Computing and Computer Assisted Intervention, pages 546--556, 2024

work page 2024

[5] [5]

Generating radiology reports via memory-driven transformer

Zhihong Chen, Yan Song, Tsung-Hui Chang, and Xiang Wan. Generating radiology reports via memory-driven transformer. In Proceedings of Conference on Empirical Methods in Natural Language Processing, pages 1439--1449, 2020

work page 2020

[6] [6]

Cross-modal memory networks for radiology report generation

Zhihong Chen, Yaling Shen, Yan Song, and Xiang Wan. Cross-modal memory networks for radiology report generation. In Proceedings of Annual Meeting of the Association for Computational Linguistics, pages 5904--5914, 2021

work page 2021

[7] [7]

Meshed-memory transformer for image captioning

Marcella Cornia, Matteo Stefanini, Lorenzo Baraldi, and Rita Cucchiara. Meshed-memory transformer for image captioning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10578--10587, 2020

work page 2020

[8] [8]

Anatomic pathology quality assurance through peer review

John Doe and Alice Smith. Anatomic pathology quality assurance through peer review. PathologyOutlines, 2023. https://www.pathologyoutlines.com/topic/managementlabAPQA.html

work page 2023