Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Rate-My-LoRA: Efficient and Adaptive Federated Model Tuning for Cardiac MRI Segmentation

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Rate-My-LoRA lifts cardiac MRI Dice, cuts bandwidth 94%

desk verdict A plausible and clearly written LoRA federated aggregation heuristic, but a single ambiguous sentence about an extra epoch of local fine-tuning could confound the central comparison. read the letter →

arxiv 2501.03223 v1 pith:YETZRSRC submitted 2025-01-06 cs.CV cs.DCcs.LG

Xiaoxiao He , Haizhou Shi , Ligong Han , Chaowei Tan , Bo Liu , Zihao Xu , Meng Ye , Leon Axel
show 2 more authors
Kang Li Dimitris Metaxas
This is my paper · ORCID
classification cs.CVcs.DCcs.LG
keywords federatedlearninglow-rankadaptationLoRAcardiacMRIsegmentationdataheterogeneityadaptiveaggregationparameter-efficientfine-tuningcommunicationefficiency
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 its adaptive aggregation rule, Rate-My-LoRA, improves both in-client and cross-client segmentation accuracy in LoRA-based federated learning for cardiac MRI, while slashing communication cost. The method trains a low-rank adapter on each hospital's data, merges them with weights that are adjusted whenever a client's validation accuracy drops, and reports Dice scores of 0.910 (in-client) and 0.904 (cross-client), outperforming FedPETuning and full-weight FedAvg. If correct, this would let hospitals with limited bandwidth pool their data for segmentation without sharing images, counteracting the performance loss caused by non-IID imaging data.

What carries the argument

The load-bearing object is the adaptive weight rule $w_t(c) = 1-\lambda$ exactly when at least one client $i$ satisfies $P^i_t < P^i_{t-1}$ and client $c$ satisfies $P^c_t > P^c_{t-1}$; otherwise $w_t(c)=1$. These weights enter the aggregation $W_t = W_{t-1} + \frac{1}{\sum_c |D_c|} \sum_c w_t(c) |D_c| B^c_t A^c_t$, where $B^c_t A^c_t$ is the LoRA adapter from client $c$. The diminishing $\lambda$ schedule makes the method converge to FedAvg as $\lambda \to 0$. The LoRA decomposition $W = W_0 + BA$ with small rank $r$ is what keeps communication to a few megabytes per round.

What would settle it

A head-to-head comparison that applies an identical one-epoch local fine-tuning step to FedPETuning and full-weight FedAvg, or removes it from Rate-My-LoRA, would directly test whether the adaptive weighting is responsible for the Dice improvement. If the gap disappears under either condition, the central claim is not supported.

Watch

Extended reading notes

Core claim

Rate-My-LoRA's central claim is that evaluating the merged model on each client's local validation set and using the resulting accuracy changes to penalize non-generalizing adapters yields a better aggregated model than dataset-size-based weighting. After each round, the server sends the equal-weight merged model to all clients; each client reports its validation accuracy. If any client's accuracy falls relative to the previous round, the server sets the aggregation weight of every other client's adapter to $1-\lambda$, with $\lambda$ decaying by 5% per round, and uses dataset-size-weighted aggregation with these adaptive weights. This is shown to raise in-client Dice from 0.840/0.883/0.873 (FedPETuning) to 0.889/0.910/0.895 across the three clients, and cross-client Dice from 0.849/0.865/0.862 to 0.882/0.894/0.904, while transmitting only the compact LoRA adapters.

Load-bearing premise

The reported advantage rests on the assumption that the extra local fine-tuning epoch after aggregation, described in the experiments, was applied identically to every compared method; if only Rate-My-LoRA received that extra training, the observed gains could come from that step rather than from the adaptive weighting rule.

Editorial extensions

If this is right

  • Hospitals with different MRI vendors, such as GE, Philips, and Siemens, can participate in a shared cardiac segmentation model without transferring patient images, with per-round communication of 1.8–3.6 MB rather than 28 MB.
  • The adaptive weighting rule can be applied to any parameter-efficient tuning method that exchanges compact adapters, not only LoRA.
  • The reported improvement in cross-client accuracy implies that knowledge from one hospital's data can transfer to other hospitals' imaging conditions, which is the primary goal of federated learning in medical imaging.
  • The method's reliance on local validation scores ties aggregation quality to the availability of a reliable validation set at each client.

Reading between the lines

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

  • A natural next step is to test whether the validation-drop trigger can be replaced by a more robust signal, such as the magnitude of the drop or a confidence interval, to avoid penalizing adapters when validation noise causes false drops.
  • The 5% decay of $\lambda$ is a heuristic; one could adapt the decay schedule based on the variance of validation scores across rounds, which is not explored in the paper.
  • Because the method explicitly compares validation accuracy across rounds, it assumes clients have enough local validation data to produce stable measurements; in few-shot settings with tiny validation sets, the penalty trigger could become overly sensitive.
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

4 major / 5 minor

Summary. The paper proposes Rate-My-LoRA, a federated learning method for cardiac MRI segmentation that combines low-rank adaptation (LoRA) with an adaptive aggregation rule. After local training of LoRA adapters, each client evaluates the preliminarily merged model on its local validation set and reports the accuracy; the server then applies a penalty weight (1−λ) to adapters according to Eq. (1) to mitigate data heterogeneity. Experiments on the M&Ms-2 dataset with an ACDC-pretrained U-Net report improved in-client and cross-client Dice scores over Local Only, full-weight FedAvg, and FedPETuning, while reducing communication bandwidth by up to 94%.

Significance. If the empirical results are trustworthy, the paper makes a useful practical contribution: a simple, communication-efficient FL scheme for heterogeneous medical image segmentation, with a plausible mechanism (validation-guided penalization) that could translate to other domains. The bandwidth savings are clearly quantified, and the paper provides qualitative visualizations as supporting evidence. However, the significance is tempered by the fragility of the experimental comparison, which hinges on an ambiguous training protocol, and by missing implementation details that would be needed for reproduction.

major comments (4)
  1. [Section 3, Experimental Settings] The sentence 'After aggregation, models are fine-tuned on local data for one extra epoch' is ambiguous: it does not explicitly state whether this extra epoch was applied to all baselines (Local Only, Average Weighting, FedPETuning) or only to Rate-My-LoRA. Because one extra epoch of local fine-tuning is a known performance booster under domain shift, the reported gains in Tables 1 and 2 (e.g., client 2 Dice 0.910 vs 0.883 for FedPETuning) could be entirely due to this extra training step rather than to the aggregation rule in Eq. (1). Please clarify the protocol and add ablations: (a) Rate-My-LoRA without the extra epoch, and (b) each baseline with the extra epoch, so that the contribution of the adaptive weighting is isolated.
  2. [Section 2, Eq. (1)] The adaptive weighting rule is written as w_t(c)=1−λ only if there exists a client i with a validation accuracy drop while client c's accuracy improves. The accompanying text and Fig. 4 describe a different-sounding rule: when the aggregated model causes a performance drop on client 3, the adapters of the *other* clients (1 and 2) are penalized. The equation is consistent with the figure only if the non-dropping clients simultaneously improve, which is not stated or guaranteed. As written, if multiple clients drop and only one improves, only the improving client is penalized, which may frustrate the stated goal of penalizing adapters that reduce generalizability. Please clarify the intended rule and ensure Eq. (1), the text, and Fig. 4 are mutually consistent.
  3. [Section 2, Eq. (2) and Algorithm 1] The aggregation formula and algorithm are underspecified. Eq. (2) writes W_t = W_{t-1} + (1/Σ|D_c|) Σ w_t(c)|D_c| B_c^t A_c^t, but it is not clear whether B_c^t A_c^t denotes the full adapter weights after local training or a per-round delta, nor how W_t relates to the frozen pretrained model W0 and the adapters from previous rounds. Algorithm 1's TrainLocal call similarly mixes the global model and all adapters without specifying which parameters are updated. Without a precise definition of these quantities, the method is not reproducible.
  4. [Section 3, Experimental Settings] The paper does not report standard deviations or significance tests for the Dice/VOE/HD/ASSD results. The improvements in Tables 1 and 2 are presented as point estimates from (apparently) a single run, so it is unclear whether the observed gaps, some of which are modest, are within run-to-run variability. Please report results over multiple random seeds or provide confidence intervals, especially for the central comparison against FedPETuning.
minor comments (5)
  1. [Section 3, Experimental Settings] LoRA is normally defined for linear layers, but the paper says 'LoRA adapters injected into each convolution block' without describing the reformulation for convolutions. Please provide a brief explanation or cite the implementation approach used.
  2. [Table 1] The Hausdorff distance values for 'Local Only' (e.g., 89.897, 93.124) are an order of magnitude larger than those of other methods; please verify the units or check for a typo, as these values seem implausible for cardiac segmentation.
  3. [Section 2, Eq. (1)] The condition in Eq. (1) uses strict inequality (P_c^t > P_c^{t-1}); if a client's accuracy stays flat while another drops, that client's adapter is not penalized. Consider whether ties should also trigger the penalty for consistency with the described behavior.
  4. [Experimental Settings] Several standard training hyperparameters are missing, including the number of communication rounds T, batch size, learning rate, optimizer, and total training epochs. Please report these to enable reproducibility.
  5. [Abstract and Conclusion] The abstract states 'up to 94% communication bandwidth' while Section 3 says 'saves up to 15.5x'; please make the relationship explicit (15.5x corresponds to ~93.5% reduction) to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the adaptive aggregation scheme is an empirical heuristic whose reported metrics are computed on disjoint test sets, and the LoRA/FL foundations rest on external citations.

full rationale

The paper's central claim is empirical: the Rate-My-LoRA aggregation rule in Eq. (1)-(2) improves in-client and cross-client Dice over LoRA-based baselines. The method uses each client's local validation accuracy P_c^t to set adaptive weights w_t(c). This is a model-selection heuristic, not a fitted quantity renamed as a prediction: the train/validation/test split is explicitly 8:1:1, so the validation Dice used to choose weights is not the test Dice reported in Tables 1 and 2. No equation defines its output in terms of its own inputs by construction. The LoRA foundation is externally cited (Hu et al., Ref. [17]); the few-shot generalization citations include external Refs. [18,20] alongside self-citations, so no load-bearing self-citation chain is present. The diffusion-model analogy (Refs. [23-25]) is motivational and not used to justify the method's correctness. The relation to prior performance-weighting work (Stripelis et al., Ref. [15]) is a clear external comparison, and the binary penalty rule differs from that prior weighting scheme, so this is not a renaming of a known result. The 'one extra epoch' of local fine-tuning after aggregation could be an experimental confound if not applied to all baselines, but that is an experimental-design concern, not a circularity of the derivation. Overall, the derivation chain is self-contained and empirical; no circular step can be exhibited from the paper's equations or citations.

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

No new physical or mathematical entities are introduced; the contribution is a heuristic aggregation rule with hand-set hyperparameters and an unproven proxy assumption.

free parameters (4)
  • lambda (penalty weight) = 0.2
    Initial penalty strength in Eq. (1), set by hand; no sensitivity analysis provided.
  • lambda decay factor = 0.95 per round
    Diminishing schedule in Algorithm 1 line 13; chosen ad hoc, no convergence analysis.
  • LoRA ranks per client = 16, 32, 32
    Adapter sizes chosen based on local dataset size; not tuned or justified.
  • Number of communication rounds T = not stated
    Training length is unspecified, so reported results cannot be placed in a budget context.
assumptions (4)
  • domain assumption Low-rank adaptation W = W0 + BA with r << min(d,k) provides regularization that improves generalization in few-shot federated scenarios.
    Invoked in Methods and supported by self-cited refs [18,20,26,27]; not proven for this setting.
  • ad hoc to paper Eq. (1): if some client's validation accuracy drops while another client's improves, the improving client's adapter is penalized by 1-lambda.
    This penalty rule is the core novelty and is asserted without derivation or theoretical justification.
  • ad hoc to paper Validation accuracy on each client's local validation set is a reliable proxy for adapter generalizability to other clients.
    The entire weighting scheme rests on this proxy; no evidence is given that validation accuracy predicts cross-client generalization.
  • ad hoc to paper Client adapters are trained independently with other adapters fixed, so the aggregated model combines compatible updates.
    The local training protocol is not fully specified; the algorithm implies each client updates only its own adapter, but this is not stated explicitly.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Rate-My-LoRA: Efficient and Adaptive Federated Model Tuning for Cardiac MRI Segmentation." pith.science (2026). https://pith.science/paper/YETZRSRC

@misc{pith2026250103223,
  author       = {Pith},
  title        = {Pith review of: Rate-My-LoRA: Efficient and Adaptive Federated Model Tuning for Cardiac MRI Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YETZRSRC}},
  note         = {Machine review of arXiv:2501.03223}
}
read the original abstract

Cardiovascular disease (CVD) and cardiac dyssynchrony are major public health problems in the United States. Precise cardiac image segmentation is crucial for extracting quantitative measures that help categorize cardiac dyssynchrony. However, achieving high accuracy often depends on centralizing large datasets from different hospitals, which can be challenging due to privacy concerns. To solve this problem, Federated Learning (FL) is proposed to enable decentralized model training on such data without exchanging sensitive information. However, bandwidth limitations and data heterogeneity remain as significant challenges in conventional FL algorithms. In this paper, we propose a novel efficient and adaptive federate learning method for cardiac segmentation that improves model performance while reducing the bandwidth requirement. Our method leverages the low-rank adaptation (LoRA) to regularize model weight update and reduce communication overhead. We also propose a \mymethod{} aggregation technique to address data heterogeneity among clients. This technique adaptively penalizes the aggregated weights from different clients by comparing the validation accuracy in each client, allowing better generalization performance and fast local adaptation. In-client and cross-client evaluations on public cardiac MR datasets demonstrate the superiority of our method over other LoRA-based federate learning approaches.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 30 canonical work pages

  1. [1]

    Rate-My-LoRA: Efficient and Adaptive Federated Model Tuning for Cardiac MRI Segmentation

    INTRODUCTION Cardiac dyssynchrony, where the heart’s ventricles beat out of sync, worsens heart failure prognosis, with up to 60% of patients dying within four years [1]. Advances in diagnostic tools like cardiac MRI (cMRI) are crucial in addressing this issue [2]. Accurate segmentation of the left ventricle cavity (LVC), myocardium (LVM), and right ventr...

    work page Pith review arXiv 2025

  2. [2]

    3 provides a visual representation of our proposed LoRA- based federated learning approach

    METHODS Fig. 3 provides a visual representation of our proposed LoRA- based federated learning approach. In essence, we fine-tune a low-rank adapter within each client using local datasets and subsequently merge these adapters using our novel Rate-My- LoRA aggregation method. Problem formulation. We consider a scenario with |C| clients, where each client ...

    work page

  3. [3]

    Client: Receive preliminarymerged adapters and evaluate

    Client: Get merged adapters from server, train locally and upload local adapter2. Client: Receive preliminarymerged adapters and evaluate

    work page

  4. [4]

    𝐴"The aggregated model 𝑊!#

    Server: Based on client evaluation, update adapter weights with 𝜆 and distribute adapters. 𝑤!𝑤"𝑤# Weight Frozen Fine-tune Adapter 1Fine-tune Adapter 2Fine-tune Adapter 31 2 3 Communication Evaluation Weight FrozenWeight Frozen 🔥 🔥 ❄Frozen ❄Frozen 🔥Fine-Tune Fig. 3. Overview of our efficient and adaptive FL method: the server contains a pretrained model an...

    work page

  5. [5]

    EXPERIMENTS Experimental Settings. This paper uses two publicly avail- able datasets: the Automated Cardiac Diagnosis Challenge (ACDC) [28] and the Multi-Disease, Multi-View & Multi- Center Right Ventricular Segmentation in Cardiac MRI (M&Ms-2) [29, 30]. The ACDC dataset, with 100 patients imaged using a Siemens scanner, is used to train the base model. T...

    work page

  6. [6]

    Rate-My-LoRA carefully evaluates the generalizability of LoRA adapters from different clients and enables fast local adaptation

    CONCLUSION AND DISCUSSION In this paper, we proposed a FL-based low-rank adaptation method, Rate-My-LoRA, to improve both in-client and cross- client cardiac segmentation accuracy under communication bandwidth constraints. Rate-My-LoRA carefully evaluates the generalizability of LoRA adapters from different clients and enables fast local adaptation. Both ...

    work page

  7. [7]

    Eth- ical approval was not required as confirmed by the license attached with the open access data

    COMPLIANCE WITH ETHICAL STANDARDS This research study was conducted retrospectively using hu- man subject data made available in open access by Universitat de Barcelona, Spain and University of Lyon, France. Eth- ical approval was not required as confirmed by the license attached with the open access data

    work page

  8. [8]

    Metaxas through NSF: IUCRC CARTA 1747778, 2235405, 2212301, 1951890, 2003874, and NIH-5R01HL127661

    ACKNOWLEDGMENTS This research has been partially funded by research grants to D. Metaxas through NSF: IUCRC CARTA 1747778, 2235405, 2212301, 1951890, 2003874, and NIH-5R01HL127661

    work page

Show all 39 references
  1. [9]

    The prog- nostic impact of dynamic ventricular dyssynchrony in patients with idiopathic dilated cardiomyopathy and nar- row QRS,

    Antonello D’Andrea, Donato Mele, et al., “The prog- nostic impact of dynamic ventricular dyssynchrony in patients with idiopathic dilated cardiomyopathy and nar- row QRS,” European Heart Journal - Cardiovascular Imaging, vol. 14, no. 2, pp. 183–189, 08 2012

    2012

  2. [10]

    Clinical applications of cardiovascular magnetic resonance imaging,

    Constantin B Marcu, Aernout M Beek, and Albert C Van Rossum, “Clinical applications of cardiovascular magnetic resonance imaging,” Cmaj, vol. 175, no. 8, pp. 911–917, 2006

    2006

  3. [11]

    Dmcvr: Morphology-guided diffusion model for 3d cardiac vol- ume reconstruction,

    Xiaoxiao He, Chaowei Tan, Ligong Han, Bo Liu, Leon Axel, Kang Li, and Dimitris N Metaxas, “Dmcvr: Morphology-guided diffusion model for 3d cardiac vol- ume reconstruction,” in International Conference on Medical Image Computing and Computer-Assisted In- tervention. Springer, 2...

    2023

  4. [12]

    Effective 3d humerus and scapula extraction using low-contrast and high- shape-variability mr data,

    Xiaoxiao He, Chaowei Tan, Yuting Qiao, Virak Tan, Dimitris Metaxas, and Kang Li, “Effective 3d humerus and scapula extraction using low-contrast and high- shape-variability mr data,” in Medical Imaging 2019: Biomedical Applications in Molecular, Structural, and Functional Imag...

    2019

  5. [13]

    Region proposal rectification towards robust instance segmentation of bi- ological images,

    Qilong Zhangli, Jingru Yi, Di Liu, Xiaoxiao He, Zhaoyang Xia, Qi Chang, Ligong Han, Yunhe Gao, Song Wen, Haiming Tang, et al., “Region proposal rectification towards robust instance segmentation of bi- ological images,” in Medical Image Computing and Computer Assisted Interven...

    2022

  6. [14]

    Transfusion: multi-view divergent fusion for medical image segmentation with transformers,

    Di Liu, Yunhe Gao, Qilong Zhangli, Ligong Han, Xi- aoxiao He, Zhaoyang Xia, Song Wen, Qi Chang, Zhen- nan Yan, Mu Zhou, et al., “Transfusion: multi-view divergent fusion for medical image segmentation with transformers,” in Medical Image Computing and Com- puter Assisted Inter...

    2022

  7. [15]

    Recursive 3d segmentation of shoulder joint with coarse-scanned mr image,

    Xiaoxiao He, Chaowei Tan, Virak Tan, and Kang Li, “Recursive 3d segmentation of shoulder joint with coarse-scanned mr image,” arXiv preprint arXiv:2203.07846, 2022

    2022 arXiv

  8. [16]

    Deep learning for cardiac image segmentation: a review,

    Chen Chen, Chen Qin, Huaqi Qiu, Giacomo Tarroni, Jinming Duan, Wenjia Bai, et al., “Deep learning for cardiac image segmentation: a review,” Frontiers in Cardiovascular Medicine, vol. 7, pp. 25, 2020

    2020

  9. [17]

    Communication-efficient learning of deep networks from decentralized data,

    Brendan McMahan, Eider Moore, Daniel Ram- age, Seth Hampson, and Blaise Aguera y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics. PMLR, 2017, pp. 1273–1282

    2017

  10. [18]

    Fed- erated learning enables big data for rare cancer boundary detection,

    Sarthak Pati, Ujjwal Baid, Brandon Edwards, Micah Sheller, Shih-Han Wang, G Anthony Reina, et al., “Fed- erated learning enables big data for rare cancer boundary detection,” Nature communications, vol. 13, no. 1, pp. 7346, 2022

    2022

  11. [19]

    Dealing with heterogeneous 3d mr knee images: A federated few-shot learning method with dual knowledge distillation,

    Xiaoxiao He, Chaowei Tan, Bo Liu, Liping Si, Weiwu Yao, Liang Zhao, Di Liu, Qilong Zhangli, Qi Chang, Kang Li, et al., “Dealing with heterogeneous 3d mr knee images: A federated few-shot learning method with dual knowledge distillation,” arXiv preprint arXiv:2303.14357, 2023

    2023 arXiv

  12. [20]

    Mining multi- center heterogeneous medical data with distributed syn- thetic learning,

    Qi Chang, Zhennan Yan, Mu Zhou, Hui Qu, Xiaoxiao He, Han Zhang, Lohendran Baskaran, Subhi Al’Aref, Hongsheng Li, Shaoting Zhang, et al., “Mining multi- center heterogeneous medical data with distributed syn- thetic learning,” Nature communications, vol. 14, no. 1, pp. 5510, 2023

    2023

  13. [21]

    Fedtune: A deep dive into efficient federated fine-tuning with pre-trained transformers,

    Jinyu Chen, Wenchao Xu, Song Guo, Junxiao Wang, Jie Zhang, and Haozhao Wang, “Fedtune: A deep dive into efficient federated fine-tuning with pre-trained transformers,” arXiv preprint arXiv:2211.08025, 2022

    2022 arXiv

  14. [22]

    FedED: Federated learning via ensemble distillation for medical relation extraction,

    Dianbo Sui, Yubo Chen, Jun Zhao, Yantao Jia, Yuantao Xie, and Weijian Sun, “FedED: Federated learning via ensemble distillation for medical relation extraction,” in Proceedings of the 2020 Conference on Empirical Meth- ods in Natural Language Processing (EMNLP), Online, Nov. 2...

    2020

  15. [23]

    Performance Weighting for Robust Feder- ated Learning Against Corrupted Sources,

    Dimitris Stripelis, Marcin Abram, and Jose Luis Am- bite, “Performance Weighting for Robust Feder- ated Learning Against Corrupted Sources,” May 2022, arXiv:2205.01184 [cs]

    2022 arXiv

  16. [24]

    Client selection based weighted federated few-shot learning,

    Xinlei Xu, Saisai Niu, Zhe Wang, Dongdong Li, Hai Yang, and Wenli Du, “Client selection based weighted federated few-shot learning,” Applied Soft Computing , vol. 128, pp. 109488, Oct. 2022

    2022

  17. [25]

    Lora: Low-rank adaptation of large language models,

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, et al., “Lora: Low-rank adaptation of large language models,” arXiv preprint arXiv:2106.09685, 2021

    2021 arXiv

  18. [26]

    Rewriting a deep generative model,

    David Bau, Steven Liu, Tongzhou Wang, Jun-Yan Zhu, and Antonio Torralba, “Rewriting a deep generative model,” in Computer Vision–ECCV 2020: 16th Euro- pean Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part I 16. Springer, 2020, pp. 351–369

    2020

  19. [27]

    Blob: Bayesian low-rank adaptation by backpropagation for large language mod- els,

    Yibin Wang, Haizhou Shi, Ligong Han, Dimitris Metaxas, and Hao Wang, “Blob: Bayesian low-rank adaptation by backpropagation for large language mod- els,” arXiv preprint arXiv:2406.11675, 2024

    2024 arXiv

  20. [28]

    Asymmetry in low- rank adapters of foundation models,

    Jiacheng Zhu, Kristjan Greenewald, Kimia Nadjahi, Haitz S ´aez de Oc ´ariz Borde, Rickard Br ¨uel Gabriels- son, Leshem Choshen, et al., “Asymmetry in low- rank adapters of foundation models,” arXiv preprint arXiv:2402.16842, 2024

    2024 arXiv

  21. [29]

    Training-free bayesianization for low- rank adapters of large language models,

    Haizhou Shi, Yibin Wang, Ligong Han, Huan Zhang, and Hao Wang, “Training-free bayesianization for low- rank adapters of large language models,” arXiv preprint arXiv:2412.05723, 2024

    2024

  22. [30]

    Fedpetuning: When federated learning meets the parameter-efficient tuning methods of pre-trained language models,

    Zhuo Zhang, Yuanhang Yang, Yong Dai, Qifan Wang, Yue Yu, Lizhen Qu, and Zenglin Xu, “Fedpetuning: When federated learning meets the parameter-efficient tuning methods of pre-trained language models,” in An- nual Meeting of the Association of Computational Lin- guistics 2023. A...

    2023

  23. [31]

    Prox- edit: Improving tuning-free real image editing with proximal guidance,

    Ligong Han, Song Wen, Qi Chen, Zhixing Zhang, Kun- peng Song, Mengwei Ren, Ruijiang Gao, Anastasis Stathopoulos, Xiaoxiao He, Yuxiao Chen, et al., “Prox- edit: Improving tuning-free real image editing with proximal guidance,” in Proceedings of the IEEE/CVF Winter Conference on...

    2024

  24. [32]

    Dice: Dis- crete inversion enabling controllable editing for multi- nomial diffusion and masked generative models,

    Xiaoxiao He, Ligong Han, Quan Dao, Song Wen, Min- hao Bai, Di Liu, Han Zhang, Martin Renqiang Min, Felix Juefei-Xu, Chaowei Tan, et al., “Dice: Dis- crete inversion enabling controllable editing for multi- nomial diffusion and masked generative models,” arXiv preprint arXiv:24...

  25. [33]

    Score-guided diffusion for 3d human re- covery,

    Anastasis Stathopoulos, Ligong Han, and Dimitris Metaxas, “Score-guided diffusion for 3d human re- covery,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 906–915

    2024

  26. [34]

    Svdiff: Com- pact parameter space for diffusion fine-tuning,

    Ligong Han, Yinxiao Li, Han Zhang, Peyman Milan- far, Dimitris Metaxas, and Feng Yang, “Svdiff: Com- pact parameter space for diffusion fine-tuning,” in Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 7323–7334

    2023

  27. [35]

    Spectrum-aware param- eter efficient fine-tuning for diffusion models,

    Xinxi Zhang, Song Wen, Ligong Han, Felix Juefei-Xu, Akash Srivastava, Junzhou Huang, Hao Wang, Molei Tao, and Dimitris N Metaxas, “Spectrum-aware param- eter efficient fine-tuning for diffusion models,” arXiv preprint arXiv:2405.21050, 2024

    2024 arXiv

  28. [36]

    Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved?,

    Olivier Bernard, Alain Lalande, et al., “Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved?,” IEEE transactions on medical imaging, vol. 37, no. 11, pp. 2514–2525, 2018

    2018

  29. [37]

    Multi-centre, multi-vendor and multi- disease cardiac segmentation: the m&ms challenge,

    Victor M Campello, Polyxeni Gkontra, Cristian Izquierdo, Carlos Martin-Isla, Alireza Sojoudi, Peter M Full, et al., “Multi-centre, multi-vendor and multi- disease cardiac segmentation: the m&ms challenge,” IEEE Transactions on Medical Imaging, vol. 40, no. 12, pp. 3543–3554, 2021

    2021

  30. [38]

    Deep learning segmentation of the right ventricle in cardiac mri: The m&ms challenge,

    Carlos Mart ´ın-Isla, V´ıctor M Campello, et al., “Deep learning segmentation of the right ventricle in cardiac mri: The m&ms challenge,” IEEE Journal of Biomedi- cal and Health Informatics, 2023

    2023

  31. [39]

    U-net: Convolutional networks for biomedical im- age segmentation,

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox, “U-net: Convolutional networks for biomedical im- age segmentation,” in Medical image computing and computer-assisted intervention–MICCAI 2015: 18th in- ternational conference, Munich, Germany, October 5-9, 2015, proceedings,...

    2015

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.