EnTrust: Modeling Inter-Modal Conflict for Trustworthy Multimodal Medical Image Analysis

Abhijit Das; Behzad Bozorgtabar; Deepak Nayak; Dwarikanath Mahapatra; Imran Razzak; Mauricio Reyes; Sudipta Roy; Zongyuan Ge

arxiv: 2606.21384 · v1 · pith:K2W76A3Znew · submitted 2026-06-19 · 💻 cs.CV · cs.AI

EnTrust: Modeling Inter-Modal Conflict for Trustworthy Multimodal Medical Image Analysis

Dwarikanath Mahapatra , Abhijit Das , Behzad Bozorgtabar , Zongyuan Ge , Sudipta Roy , Deepak Nayak , Mauricio Reyes , Imran Razzak This is my paper

Pith reviewed 2026-06-26 14:27 UTC · model grok-4.3

classification 💻 cs.CV cs.AI

keywords multimodal medical imaginginter-modal conflictuncertainty estimationdiffusion modelsimage segmentationcalibrationtrustworthy AIfeature disentanglement

0 comments

The pith

EnTrust decomposes multimodal medical images into consensus, specific, and conflict features to tie uncertainty directly to modal disagreements.

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

The paper aims to show that explicitly modeling inter-modal conflict improves both segmentation accuracy and the trustworthiness of uncertainty estimates in multimodal medical imaging. Current approaches either average over disagreements or estimate uncertainty separately from the fusion process. By disentangling features into shared, modality-specific, and conflict components, and conditioning a diffusion model on this decomposition, EnTrust produces predictions that diverge where modalities disagree. This leads to better calibrated uncertainty maps that explain why a prediction is unreliable. The approach achieves state-of-the-art results on multiple benchmarks while using fewer resources than ensemble methods.

Core claim

EnTrust introduces an EnFuse module that decomposes multimodal features into shared anatomical consensus (F_c), modality-specific cues (F_{u,m}), and spatially localized conflict signals (F_{cf}), enforcing independence via cross-covariance. This decomposition conditions a diffusion-based segmentation model (SegDiff) so that sampled hypotheses diverge in regions of modal disagreement. A TrustMap module then converts this divergence into pixel-wise calibrated uncertainty, allowing clinicians to see not only where but why predictions are uncertain.

What carries the argument

The EnFuse module, which disentangles multimodal features into three independent components (shared consensus, modality-specific, and conflict) using a cross-covariance objective to condition the diffusion segmentation model.

If this is right

Segmentation accuracy reaches state-of-the-art levels across brain, cardiac, lesion, and oncology benchmarks.
Calibration error drops by 40% compared to the strongest baseline.
A single model outperforms five times deep ensembles while using roughly half the memory footprint.
Uncertainty estimates become interpretable by linking them directly to inter-modal conflicts.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the decomposition works as claimed, the method could extend to other multimodal fusion tasks beyond segmentation, such as registration or classification.
Testing on datasets with more than two modalities might reveal whether the three-component split scales or needs adjustment.
The reduced memory footprint suggests potential for deployment in resource-constrained clinical settings where ensembles are impractical.

Load-bearing premise

The cross-covariance objective enforces statistical independence among the shared, specific, and conflict feature components.

What would settle it

On a held-out multimodal dataset, if the sampled diffusion hypotheses do not show increased divergence specifically in regions where the input modalities disagree on anatomy or pathology, the central mechanism would be falsified.

Figures

Figures reproduced from arXiv: 2606.21384 by Abhijit Das, Behzad Bozorgtabar, Deepak Nayak, Dwarikanath Mahapatra, Imran Razzak, Mauricio Reyes, Sudipta Roy, Zongyuan Ge.

**Figure 1.** Figure 1: EnFuse disentanglement on (a) real BraTS 2021 MRI and (b) synthetic phantoms. Fc captures modality-invariant anatomy, Fu,T1ce isolates T1ce-specific contrast enhancement, and Fcf localizes inter-modal disagreement at tumor boundaries. obstacle to clinical deployment [6,8]. Radiologists routinely fuse complementary modalities (MRI sequences, CT, PET) to reach diagnostic judgments, yet these modalities freq… view at source ↗

**Figure 2.** Figure 2: EnTrust architecture. ModEnc extracts per-modality features, which EnFuse disentangles into shared (Fc), unique (Fu,m), and conflict (Fcf ) components. The gated fusion Fen conditions SegDiff for probabilistic mask generation. TrustMap derives pixelwise uncertainty via ensemble entropy H(p), ProbeStep Yˆ K 0 , and CalibHead C(p). before fusion. Ec, Eu,m, and Dcf are lightweight three-block ConvNeXt heads:… view at source ↗

**Figure 3.** Figure 3: Qualitative comparison on BraTS 2021. EnTrust produces boundarylocalized uncertainty aligned with Fcf ; baselines exhibit progressively diffuse estimates. U-Net P-UNet Deep Ens.(5×) DEviS MedSeg Diff EnTrust (Ours) (a) Training Cost & Model Size 0 10 20 30 40 50 60 70 80 Trainin g G P U-h o urs (4 × A10 0) 14h 20h 70h 18h 48h 58h 0 25 50 75 100 125 150 175 200 Parameters (M) 35M 42M 175M 38M 48M 62M Param… view at source ↗

**Figure 4.** Figure 4: Computational analysis. (a) Training cost and model size. (b) Accuracy vs. inference cost (bubble = peak GPU memory). (c) Inference memory breakdown. (d) UQ quality comparison. role at conflict boundaries. The encoder modularity block (blue rows) demonstrates that swapping Swin-T for ResNet-50 or ViT-B/16 without modifying EnFuse or TrustMap yields competitive results, confirming encoder-agnostic design.… view at source ↗

read the original abstract

Multimodal medical imaging fuses complementary anatomical and functional information, yet modalities frequently disagree in pathologically heterogeneous regions. Current segmentation models handle this in one of two inadequate ways: deterministic fusion that averages away disagreement, or post-hoc uncertainty estimation decoupled from the fusion process that produces it. Both obscure the clinically critical question: why is this prediction unreliable? We present EnTrust, a framework that treats inter-modal conflict as the primary source of predictive uncertainty. Our EnFuse module decomposes multimodal features into three disentangled components: shared anatomical consensus (F_c), modality-specific cues (F_{u,m}), and spatially localized conflict signals (F_{cf}), with independence enforced via a cross-covariance objective. This structured decomposition conditions SegDiff, a diffusion-based generative segmentation model whose sampled hypotheses diverge specifically in regions of modal disagreement. TrustMap then translates this hypothesis divergence into calibrated, pixel-wise uncertainty using ensemble entropy, conflict-guided perturbation probing, and a learned calibration head, enabling clinicians to understand not only where predictions are uncertain, but why. Across four benchmarks spanning brain, cardiac, lesion, and oncology domains, EnTrust achieves state-of-the-art segmentation accuracy while reducing calibration error by 40% compared to the strongest baseline. Notably, it outperforms 5x deep ensembles using a single model at roughly half the memory footprint. Code and checkpoints are available at https://github.com/GenMI-Lab/EnTrust.git.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

EnTrust decomposes features into consensus, modality-specific, and conflict parts then conditions diffusion on the conflict signal to make uncertainty reflect disagreement, but cross-covariance does not guarantee the required independence.

read the letter

The main point is that this paper builds a pipeline where multimodal features are split into three parts, the conflict part conditions a diffusion segmentation model, and the resulting hypothesis spread is turned into a calibrated uncertainty map. That structure directly addresses why a prediction is uncertain rather than just reporting that it is.

The decomposition step and the conditioning of SegDiff on conflict features are the actual new pieces. Prior work either fuses deterministically or adds uncertainty afterward; here the generative process is steered by the disagreement signal. The clinical motivation is clear and the memory claim versus five-model ensembles is worth checking if the numbers hold.

The soft spot is the cross-covariance loss used to enforce independence among the three feature groups. It only removes linear second-order correlations. Medical image features extracted by CNNs routinely show higher-order dependence, so residual statistical links between the components are likely. If that happens, the diffusion samples will not diverge specifically where modalities disagree, and TrustMap uncertainty will not be cleanly conflict-driven. The abstract performance numbers (SOTA accuracy, 40% calibration drop) cannot be assessed without the tables, but the independence gap is structural rather than minor.

This is for researchers working on uncertainty in multimodal medical segmentation who want a mechanism that ties the two together. A reader focused on clinical deployment or calibration would find the framing useful even if the implementation needs tightening. The paper deserves a serious referee because the problem is real and the proposed link between conflict and uncertainty is coherent, though the independence claim needs direct examination.

Referee Report

1 major / 2 minor

Summary. The manuscript proposes EnTrust, a framework for multimodal medical image segmentation that treats inter-modal conflict as the primary source of uncertainty. The EnFuse module decomposes input features into shared consensus (F_c), modality-specific (F_{u,m}), and conflict (F_cf) components, with independence enforced by a cross-covariance objective. These condition SegDiff, a diffusion-based generative model, so that sampled hypotheses diverge specifically in regions of modal disagreement. TrustMap converts the resulting hypothesis divergence into pixel-wise calibrated uncertainty via ensemble entropy, conflict-guided perturbation, and a learned calibration head. The paper reports state-of-the-art segmentation accuracy and a 40% reduction in calibration error versus the strongest baseline across four benchmarks (brain, cardiac, lesion, oncology), while outperforming 5× deep ensembles with a single model at roughly half the memory cost. Code is released.

Significance. If the feature disentanglement isolates conflict-driven uncertainty as claimed, the approach would provide a clinically useful link between modal disagreement and predictive unreliability, advancing trustworthy multimodal fusion beyond post-hoc uncertainty methods. The reported efficiency gain over ensembles would be a practical advantage if substantiated. Public code and checkpoints strengthen reproducibility.

major comments (1)

[EnFuse module (cross-covariance objective)] EnFuse module (cross-covariance objective): the loss is presented as enforcing statistical independence among F_c, F_{u,m}, and F_cf so that SegDiff hypotheses diverge specifically where modalities disagree. Cross-covariance penalizes only linear second-order correlations and does not guarantee full statistical independence for the non-Gaussian, higher-order dependent features typical of CNN-extracted medical images. This assumption is load-bearing for the central claim that TrustMap uncertainty is conflict-driven rather than generic.

minor comments (2)

[Method section] Notation for modality-specific features is written F_{u,m}; clarify whether the subscript denotes per-modality or per-sample indexing and ensure consistency across equations and figures.
[Abstract] The abstract states quantitative improvements (SOTA accuracy, 40% calibration reduction) without citing the supporting tables or figures; add explicit references in the abstract or introduction.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for highlighting this important technical point on the EnFuse module. We address the concern directly below and propose a targeted revision.

read point-by-point responses

Referee: EnFuse module (cross-covariance objective): the loss is presented as enforcing statistical independence among F_c, F_{u,m}, and F_cf so that SegDiff hypotheses diverge specifically where modalities disagree. Cross-covariance penalizes only linear second-order correlations and does not guarantee full statistical independence for the non-Gaussian, higher-order dependent features typical of CNN-extracted medical images. This assumption is load-bearing for the central claim that TrustMap uncertainty is conflict-driven rather than generic.

Authors: We agree that the cross-covariance objective enforces only linear second-order decorrelation and cannot guarantee full statistical independence for the non-Gaussian, higher-order dependencies present in CNN features. This is a substantive limitation of the current formulation. We will revise the manuscript to (1) explicitly describe the objective as promoting linear independence rather than claiming full statistical independence, (2) add a dedicated limitations paragraph discussing the approximation and its potential impact on the conflict-driven uncertainty claim, and (3) include an ablation study comparing cross-covariance against a mutual-information estimator (e.g., via MINE) on one benchmark to quantify the practical difference. These changes will make the load-bearing assumption transparent while preserving the empirical evidence that the current objective yields conflict-aligned uncertainty maps. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation uses explicit losses and generative sampling without reduction to inputs

full rationale

The paper defines feature decomposition via an explicit cross-covariance objective, conditions a diffusion model on the resulting conflict component, and derives uncertainty from hypothesis divergence plus ensemble entropy. These are forward-defined mechanisms with stated loss terms and sampling procedures rather than parameters fitted to target quantities and renamed as predictions, self-definitional loops, or load-bearing self-citations. The chain remains independent of its outputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 3 invented entities

Abstract-only review limits visibility into parameters and assumptions; the described method introduces new modules whose independence and effectiveness rest on unverified domain assumptions.

axioms (2)

domain assumption Cross-covariance objective enforces independence of consensus, modality-specific, and conflict feature components
Stated directly in the abstract description of EnFuse
domain assumption Diffusion model hypotheses diverge specifically in modal disagreement regions when conditioned on conflict signals
Abstract states this as the conditioning mechanism for SegDiff

invented entities (3)

EnFuse module no independent evidence
purpose: Decomposes multimodal features into three disentangled components
New module introduced in the framework
SegDiff no independent evidence
purpose: Diffusion-based generative segmentation model
New diffusion model variant
TrustMap no independent evidence
purpose: Maps hypothesis divergence to calibrated pixel-wise uncertainty
New uncertainty translation component

pith-pipeline@v0.9.1-grok · 5817 in / 1439 out tokens · 26266 ms · 2026-06-26T14:27:17.125978+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

17 extracted references · 1 canonical work pages

[1]

Authors affiliated with IQVIA

Arias, A., Rickwood, S.: Boosting healthcare capacity with AI (Jan 2024),https://www.weforum.org/stories/2024/01/ ai-in-healthcare-could-bridge-a-significant-capacity-gap/, published as part of the World Economic Forum Annual Meeting 2024. Authors affiliated with IQVIA

2024
[2]

In: International conference on medical image computing and computer-assisted intervention

Das, A., Gorade, V., Kumar, K., Chakraborty, S., Mahapatra, D., Roy, S.: Confidence-guided semi-supervised learning for generalized lesion localization in x-ray images. In: International conference on medical image computing and computer-assisted intervention. pp. 242–252. Springer (2024)

2024
[3]

Frontiers in Neuroinformat- ics13, 58 (2019) 10 Mahapatra et al

Fedorov, A., Li, X., Sarah, A.S., Kogan, Y., et al.: Atlas: A dataset for traumatic brain injury lesion segmentation and characterization. Frontiers in Neuroinformat- ics13, 58 (2019) 10 Mahapatra et al

2019
[4]

Gal, Y., Ghahramani, Z.: Dropout as a bayesian approximation: Representing modeluncertaintyindeeplearning.In:InternationalConferenceonMachineLearn- ing (ICML). pp. 1050–1059 (2016)

2016
[5]

In: Advances in Neural Information Processing Systems

Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems. vol. 33, pp. 6840–6851 (2020)

2020
[6]

Survey of 8,584 IT professionals across 10 countries

IBM: Data suggests growth in enterprise adoption of AI is due to widespread deployment by early adopters (Jan 2024),https://newsroom.ibm.com/ 2024-01-10-Data-Suggests-Growth-in-Enterprise-Adoption-of-AI-is-Due-\ to-Widespread-Deployment-by-Early-Adopters, IBM Global AI Adoption In- dex 2023, conducted by Morning Consult on behalf of IBM. Survey of 8,584 I...

2024
[7]

Kohl, S.A.A., Romera-Paredes, B., Meyer, C., Fauw, J.D., Ledsam, J.R., Maier- Hein, K.H., Eslami, S.M.A., Rezende, D.J., Ronneberger, O.: A probabilistic u- net for segmentation of ambiguous images (2019),https://arxiv.org/abs/1806. 05034

2019
[8]

In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024

Konuk, E., Welch, R., Christiansen, F., Epstein, E., Smith, K.: A framework for assessing joint human-AI systems based on uncertainty estimation. In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024. Springer, Cham (2024)

2024
[9]

Advances in Neural Information Pro- cessing Systems30, 6402–6413 (2017)

Lakshminarayanan, B., Pritzel, A., Blundell, C.: Simple and scalable predictive uncertainty estimation using deep ensembles. Advances in Neural Information Pro- cessing Systems30, 6402–6413 (2017)

2017
[10]

Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer: Hierarchical vision transformer using shifted windows (2021),https: //arxiv.org/abs/2103.14030

Pith/arXiv arXiv 2021
[11]

Construction of correlation functions in two and three dimensions

Menze, B.H., Jakab, A., Bauer, S., Kalpathy-Cramer, J., et. al.: The multimodal brain tumor segmentation challenge (brats) 2012, 2013, 2014, 2015. IEEE Transac- tions on Medical Imaging37(1), 237–250 (2019).https://doi.org/10.1109/TMI. 2018.2818988

work page doi:10.1109/tmi 2012
[12]

In: 2016 Fourth International Confer- ence on 3D Vision (3DV)

Milletari, F., Navab, N., Ahmadi, S.A.: V-net: Fully convolutional neural networks for volumetric medical image segmentation. In: 2016 Fourth International Confer- ence on 3D Vision (3DV). pp. 565–571. IEEE (2016)

2016
[13]

Organizers: Hektor 2025 challenge,https://hecktor25.grand-challenge.org/ dataset/

2025
[14]

In: Advances in Neural Information Processing Systems

Sensoy, M., Kaplan, L., Kandemir, E.: Evidential deep learning to quantify un- certainty in image classification. In: Advances in Neural Information Processing Systems. vol. 31 (2018)

2018
[15]

In: Med- ical Imaging with Deep Learning

Wu, J., Fu, R., Fang, H., Zhang, Y., Yang, Y., Xiong, H., Liu, H., Xu, Y.: Med- SegDiff: Medical Image Segmentation with Diffusion Probabilistic Model. In: Med- ical Imaging with Deep Learning. Proceedings of Machine Learning Research, vol. 227, pp. 1623–1639 (2024),https://proceedings.mlr.press/v227/wu24a. html

2024
[16]

IEEE transactions on pattern analysis and machine intelli- gence41(12), 2933–2946 (2018)

Zhuang, X.: Multivariate mixture model for myocardial segmentation combining multi-source images. IEEE transactions on pattern analysis and machine intelli- gence41(12), 2933–2946 (2018)

2018
[17]

Zou, K., Chen, Y., Huang, L., Yuan, X., Shen, X., Wang, M., Goh, R.S.M., Liu, Y., Fu, H.: Towards reliable medical image segmentation by utilizing evidential calibrated uncertainty (2024)

2024

[1] [1]

Authors affiliated with IQVIA

Arias, A., Rickwood, S.: Boosting healthcare capacity with AI (Jan 2024),https://www.weforum.org/stories/2024/01/ ai-in-healthcare-could-bridge-a-significant-capacity-gap/, published as part of the World Economic Forum Annual Meeting 2024. Authors affiliated with IQVIA

2024

[2] [2]

In: International conference on medical image computing and computer-assisted intervention

Das, A., Gorade, V., Kumar, K., Chakraborty, S., Mahapatra, D., Roy, S.: Confidence-guided semi-supervised learning for generalized lesion localization in x-ray images. In: International conference on medical image computing and computer-assisted intervention. pp. 242–252. Springer (2024)

2024

[3] [3]

Frontiers in Neuroinformat- ics13, 58 (2019) 10 Mahapatra et al

Fedorov, A., Li, X., Sarah, A.S., Kogan, Y., et al.: Atlas: A dataset for traumatic brain injury lesion segmentation and characterization. Frontiers in Neuroinformat- ics13, 58 (2019) 10 Mahapatra et al

2019

[4] [4]

Gal, Y., Ghahramani, Z.: Dropout as a bayesian approximation: Representing modeluncertaintyindeeplearning.In:InternationalConferenceonMachineLearn- ing (ICML). pp. 1050–1059 (2016)

2016

[5] [5]

In: Advances in Neural Information Processing Systems

Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems. vol. 33, pp. 6840–6851 (2020)

2020

[6] [6]

Survey of 8,584 IT professionals across 10 countries

IBM: Data suggests growth in enterprise adoption of AI is due to widespread deployment by early adopters (Jan 2024),https://newsroom.ibm.com/ 2024-01-10-Data-Suggests-Growth-in-Enterprise-Adoption-of-AI-is-Due-\ to-Widespread-Deployment-by-Early-Adopters, IBM Global AI Adoption In- dex 2023, conducted by Morning Consult on behalf of IBM. Survey of 8,584 I...

2024

[7] [7]

Kohl, S.A.A., Romera-Paredes, B., Meyer, C., Fauw, J.D., Ledsam, J.R., Maier- Hein, K.H., Eslami, S.M.A., Rezende, D.J., Ronneberger, O.: A probabilistic u- net for segmentation of ambiguous images (2019),https://arxiv.org/abs/1806. 05034

2019

[8] [8]

In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024

Konuk, E., Welch, R., Christiansen, F., Epstein, E., Smith, K.: A framework for assessing joint human-AI systems based on uncertainty estimation. In: Proceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024. Springer, Cham (2024)

2024

[9] [9]

Advances in Neural Information Pro- cessing Systems30, 6402–6413 (2017)

Lakshminarayanan, B., Pritzel, A., Blundell, C.: Simple and scalable predictive uncertainty estimation using deep ensembles. Advances in Neural Information Pro- cessing Systems30, 6402–6413 (2017)

2017

[10] [10]

Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer: Hierarchical vision transformer using shifted windows (2021),https: //arxiv.org/abs/2103.14030

Pith/arXiv arXiv 2021

[11] [11]

Construction of correlation functions in two and three dimensions

Menze, B.H., Jakab, A., Bauer, S., Kalpathy-Cramer, J., et. al.: The multimodal brain tumor segmentation challenge (brats) 2012, 2013, 2014, 2015. IEEE Transac- tions on Medical Imaging37(1), 237–250 (2019).https://doi.org/10.1109/TMI. 2018.2818988

work page doi:10.1109/tmi 2012

[12] [12]

In: 2016 Fourth International Confer- ence on 3D Vision (3DV)

Milletari, F., Navab, N., Ahmadi, S.A.: V-net: Fully convolutional neural networks for volumetric medical image segmentation. In: 2016 Fourth International Confer- ence on 3D Vision (3DV). pp. 565–571. IEEE (2016)

2016

[13] [13]

Organizers: Hektor 2025 challenge,https://hecktor25.grand-challenge.org/ dataset/

2025

[14] [14]

In: Advances in Neural Information Processing Systems

Sensoy, M., Kaplan, L., Kandemir, E.: Evidential deep learning to quantify un- certainty in image classification. In: Advances in Neural Information Processing Systems. vol. 31 (2018)

2018

[15] [15]

In: Med- ical Imaging with Deep Learning

Wu, J., Fu, R., Fang, H., Zhang, Y., Yang, Y., Xiong, H., Liu, H., Xu, Y.: Med- SegDiff: Medical Image Segmentation with Diffusion Probabilistic Model. In: Med- ical Imaging with Deep Learning. Proceedings of Machine Learning Research, vol. 227, pp. 1623–1639 (2024),https://proceedings.mlr.press/v227/wu24a. html

2024

[16] [16]

IEEE transactions on pattern analysis and machine intelli- gence41(12), 2933–2946 (2018)

Zhuang, X.: Multivariate mixture model for myocardial segmentation combining multi-source images. IEEE transactions on pattern analysis and machine intelli- gence41(12), 2933–2946 (2018)

2018

[17] [17]

Zou, K., Chen, Y., Huang, L., Yuan, X., Shen, X., Wang, M., Goh, R.S.M., Liu, Y., Fu, H.: Towards reliable medical image segmentation by utilizing evidential calibrated uncertainty (2024)

2024