Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

From Flexibility to Manipulation: The Slippery Slope of XAI Evaluation

T0 review · 3 major / 5 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 Faithfulness evaluations of XAI methods can be flipped by changing hyperparameters that papers treat as interchangeable.

desk verdict The empirical ranking-reversal result is real and useful, but the paper's formal definitions misstate their own objective and the 'widely used' hyperparameter claim lacks the survey to back it, so it needs revision before publication. read the letter →

arxiv 2412.05592 v1 pith:JRIB3K5G submitted 2024-12-07 cs.AI

Kristoffer Wickstr{o}m , Marina Marie-Claire Höhne , Anna Hedström This is my paper · ORCID
classification cs.AI
keywords XAIevaluationfaithfulnesshyperparametersensitivityadversarialmanipulationMeanResilienceRankexplainabilityreproducibilityreliability
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

Quantitative evaluation of explainable AI lacks ground-truth labels, so faithfulness metrics—tests of how well an explanation tracks a model's behavior as input parts are removed—depend on user-specified hyperparameters such as partition size, perturbation function, and normalization. This paper tries to show that those hyperparameters are flexible enough to be an attack surface: switching among settings the literature treats as acceptable can reverse which explanation method scores best. The authors define two manipulation objectives, demonstrate them on four image datasets and several explanation methods, and propose a ranking-based mitigation called Mean Resilience Rank. The stakes are practical: if manipulation is easy, published faithfulness comparisons cannot be taken at face value.

What carries the argument

The central object is the evaluation function $F(f, x, e, a, b, c) \to s$, which maps a classifier, an input, an explanation, and three hyperparameters to a faithfulness score, together with a feasible set $A^*_a$ of hyperparameter values regarded as acceptable in the literature. The argument treats the hyperparameters as optimization variables inside $A^*_a$ rather than as a fixed protocol: intra-manipulation maximizes $F$ for a single method, and inter-manipulation maximizes the score gap between a chosen method and the others. The experiments instantiate $F$ by partitioning the input, perturbing blocks in decreasing order of attribution sum, and computing the AUC of the resulting faithfulness curve, so the manipulable choices are concretely partition size, perturbation function, and normalization. Mean Resilience Rank then averages each explanation method's rank across all configurations in $A^*_a$, making it the proposed counterweight to the manipulation.

What would settle it

Count how often each hyperparameter combination from the paper's feasible set actually appears in a representative sample of published faithfulness evaluations; if a single configuration is used in the large majority of studies, then the demonstrated manipulation would rarely transfer to real published comparisons, because the claimed flexibility would not exist in practice.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that faithfulness evaluation of black-box image classifiers is manipulable. With a fixed model, fixed explanations, and a fixed evaluation metric, merely changing the perturbation function from uniform noise to Gaussian blurring can flip the best-performing explanation method, as shown for MNIST where the winner changes from Saliency to LRP while all other settings stay fixed. The paper generalizes this into two optimization problems: intra-manipulation, which chooses a hyperparameter from a feasible set to maximize one method's score, and inter-manipulation, which chooses a hyperparameter to maximize the gap between a focus method and its competitors. Experiments across MNIST, FashionMNIST, PneumoniaMNIST, and ImageNet show score changes large enough to invert rankings or even let every method be made the winner on the same data. The paper concludes that no single-configuration faithfulness score is trustworthy and proposes Mean Resilience Rank, which averages each method's rank across all feasible hyperparameter settings, as a partial remedy.

Load-bearing premise

The load-bearing premise is that the hyperparameter values in the paper's feasible set (partition sizes 14/28/56 or 112/224/448, standard-normal/uniform/Gaussian-blur perturbations, normalization on/off) are all genuinely common in published practice, so switching among them looks like a normal choice rather than an unusual one.

Editorial extensions

If this is right

  • A faithfulness score computed under a single hyperparameter configuration is not a stable measurement of explanation quality; the same methods receive different scores under equally accepted configurations.
  • Comparative claims such as 'method A beats method B' can be reversed by the evaluator's choice of partition size, perturbation function, or normalization, so papers reporting one configuration may be reporting an artifact of those choices.
  • The intra- and inter-manipulation definitions give an audit recipe: test whether an evaluation conclusion survives across the full feasible set before trusting it.
  • Mean Resilience Rank makes the evaluation more robust, but only at the cost of many more model evaluations and with results that still depend on which hyperparameters are included in the feasible set.

Reading between the lines

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

  • Going beyond the paper: the same flexibility likely extends to other metric families (e.g. randomisation-based sanity checks and robustness metrics) that also expose user-chosen parameters, so the manipulation may be a general property of XAI evaluation rather than a quirk of faithfulness tests.
  • A second-level inference: if the feasible set itself is selected by the researcher, MRR can be gamed by choosing a feasible set that happens to favor one method, so the mitigation moves the manipulation surface rather than eliminating it.
  • A testable extension: pre-registering the hyperparameter grid before inspecting any explanations, then reporting the full distribution of scores, should reduce the scope for this type of outcome manipulation in published evaluations.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper argues that quantitative faithfulness evaluation of XAI methods is vulnerable to manipulation through hyperparameter choices. The authors formalize two attacks: intra-manipulation, which improves the score of a single explanation method, and inter-manipulation, which reorders the ranking of several methods. They demonstrate across MNIST, FashionMNIST, PneumoniaMNIST, and ImageNet that exhaustive search over 18 hyperparameter configurations (partition size, perturbation function, normalization) can change which of LRP, Saliency, and KernelSHAP is ranked best. They also propose Mean Resilience Rank (MRR), a ranking-based aggregation over the feasible hyperparameter set, as a mitigation. The empirical demonstrations do show ranking reversals, but the formal definitions contain sign errors, the feasible set is asserted rather than measured, and the results are reported without uncertainty quantification.

Significance. If the central result holds, the paper makes a valuable and provocative point: faithfulness evaluations on black-box image classifiers are not a stable basis for comparing XAI methods unless hyperparameter choices are constrained or reported comprehensively. The paper's strengths include a clear and reproducible experimental setup with linked code, a concrete feasible-set definition, and a constructive mitigation proposal in MRR. The implications are substantial because faithfulness metrics are among the most widely used quantitative evaluation tools in XAI. However, the significance is tempered by the fact that the attack's realism depends on the assertion that all 18 configurations are 'common' choices; the paper provides citations but no frequency data. The formal definitions also need correction before the method can be reproduced from the text.

major comments (3)
  1. [Section 4, Definitions 1 and 2] The optimization objectives are inconsistent with the reported results. Because the paper fixes the aggregation as AUC and states that a lower faithfulness score is better (Section 5.1, Section 8 caption), Definition 2's objective 'maximize F(f,x,em,a,b,c) - sum_{m' != m} F(f,x,em',a,b,c)' would select a configuration where the focus method has a high (bad) score and the other methods have low (good) scores, which is the opposite of 'towards LRP' or 'towards Saliency.' The reported tables, e.g. Table 4 for MNIST, show the focus method with the lowest manipulated score (LRP 37.79 vs Saliency 46.23 and KernelSHAP 50.77), indicating that the implementation minimized F(em) relative to the others. The objective should be reversed, e.g. minimize F(em) - sum_{m' != m} F(em') or maximize the sum of the non-focus scores minus the focus score. Definition 1 has the analogous problem: intra-manipulation in Table 3 reduces AUC values, so the direction of optimization should be minimization, not maximization as written. Please correct the sign errors in the definitions and confirm that the code follows the corrected objectives.
  2. [Section 7.1, Table 2, and Section 8.2] The feasible set of hyperparameters is presented as 'based on common choices in the literature' with citations [7,15,24,26,53] and [3,40,46] and [10,11,24], but no quantitative evidence is given that each of the 18 configurations is actually used by a non-negligible share of published faithfulness evaluations. The abstract and introduction claim that the manipulation uses 'hyperparameters that are widely used in the literature' and 'seemingly innocent changes'; without frequency data, the demonstrated reversals could be an artifact of an author-selected grid rather than a realistic threat. The paper explicitly references 'Appendix A' for a summary of how often each hyperparameter occurs in the manipulated set, but Appendix A is missing from this version, so the reader cannot check the plausibility of the manipulated settings. Please add the appendix or a small survey of published hyperparameter choices, or temper the generalizability claims accordingly.
  3. [Section 8, Tables 3-7] All reported scores are point estimates without confidence intervals or standard deviations. Since faithfulness scores are averaged over 100-1000 samples, the observed ranking changes and the 'large changes' claim should be accompanied by bootstrap intervals, per-sample variances, or repeated sampling to show that the differences are not due to sampling noise. This is especially important for ImageNet, which uses only 100 samples, and for the conclusion that the ranking can 'completely change' after manipulation.
minor comments (5)
  1. [Section 6, Equation (5)] The MRR formula is ambiguous: R returns integer ranks 0..M-1, but Eq. (5) divides by M, yielding values in [0,(M-1)/M], while Table 7 states 'a rank of 0 is best and 1 is worst.' Either the normalization should be by M-1, or R should return ranks 1..M, or the caption should be corrected to state the actual range.
  2. [Section 8.1] The statement 'in some cases as much as a 130% improvement' is unclear; for example, LRP on MNIST changes from 25.20 to 7.86, which is a 69% decrease. Please define the relative change convention used for the 'improvement' percentages.
  3. [Section 8.2] The sentence 'For some tasks, the evaluation outcome can be manipulated such that most of the three methods achieves the best performance' should read 'all three methods' when referring to PneumoniaMNIST, and the subject-verb agreement should be fixed.
  4. [References] References [9] and [10] are the same paper (Bhatt, Weller, and Moura, IJCAI 2020) and should be merged or renumbered.
  5. [Section 5.1] The phrase 'course faithfulness curves' appears to be a typo for 'coarse faithfulness curves'.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: the manipulation results are the designed output of an explicit optimization; the feasible-set premise raises external-validity, not circularity, concerns.

full rationale

The paper's central claim is that faithfulness evaluation outcomes can be changed by selecting hyperparameters from an accepted feasible set. This is not a derivation that hides its own conclusion: Definition 1 explicitly defines intra-manipulation as 'maximize_a F(f, x, e, a, b, c) subject to a in A*_a', and Definition 2 similarly optimizes an inter-method objective. The reported improvements in Tables 3-6 are the result of running this stated search, so they are consequences of the objective, not a fitted parameter renamed as a prediction. The only external premise that could be questioned is whether the values in Table 2 are actually 'common choices'; the paper supports this with citations [7,15,24,26,53], [3,40,46], and [10,11,24], and some of these involve the authors. This is a plausibility/generalizability concern, not a circular reduction: the feasible set is an input assumption, and the manipulation experiments are meaningful only if that assumption approximates field practice. The paper itself flags this in Sec. 9 ('MRR requires domain expertise to determine the feasible set of hyperparameters. If the selection of the feasible set is done incorrectly, it might exacerbate the problem of manipulation'), and it notes that Appendix A should provide counts of hyperparameter use, though that appendix is absent from this version. Such missing evidence affects empirical support, not circularity. MRR is presented as a proposal and evaluated on the same grid; it is not a first-principles derivation from self-cited uniqueness theorems. The sign convention in Definition 2 is inconsistent with lower-is-better AUC (a correctness issue), but it does not create a circular equivalence. No step of the argument reduces by construction to its input, so the circularity burden is low.

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

No fitted parameters or invented entities are introduced. The manipulation operates on existing evaluation hyperparameters. The main assumptions are domain assumptions about XAI evaluation and the representativeness of the tested settings.

assumptions (4)
  • domain assumption There is no ground truth explanation label for the evaluated tasks.
    This premise is stated in Sec 1 and Sec 6; without ground truth, there is no objective way to choose among hyperparameters, which enables the manipulation.
  • domain assumption The chosen feasible hyperparameter set (Table 2) consists of commonly used, acceptable choices.
    The 'seemingly innocent' claim depends on these being settings the community would accept; the paper cites a few references but provides no systematic survey of usage.
  • domain assumption Faithfulness, computed as AUC of the prediction-drop curve, is a valid measure of explanation quality.
    All experiments and the MRR proposal are evaluated within this framework; if this metric is not meaningful, the vulnerability is less consequential.
  • domain assumption LeNet and ResNet18 on standard image datasets are representative of XAI use cases.
    The paper only tests two architectures and four datasets; the generality of the manipulation is assumed rather than demonstrated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Flexibility to Manipulation: The Slippery Slope of XAI Evaluation." pith.science (2026). https://pith.science/paper/JRIB3K5G

@misc{pith2026241205592,
  author       = {Pith},
  title        = {Pith review of: From Flexibility to Manipulation: The Slippery Slope of XAI Evaluation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JRIB3K5G}},
  note         = {Machine review of arXiv:2412.05592}
}
read the original abstract

The lack of ground truth explanation labels is a fundamental challenge for quantitative evaluation in explainable artificial intelligence (XAI). This challenge becomes especially problematic when evaluation methods have numerous hyperparameters that must be specified by the user, as there is no ground truth to determine an optimal hyperparameter selection. It is typically not feasible to do an exhaustive search of hyperparameters so researchers typically make a normative choice based on similar studies in the literature, which provides great flexibility for the user. In this work, we illustrate how this flexibility can be exploited to manipulate the evaluation outcome. We frame this manipulation as an adversarial attack on the evaluation where seemingly innocent changes in hyperparameter setting significantly influence the evaluation outcome. We demonstrate the effectiveness of our manipulation across several datasets with large changes in evaluation outcomes across several explanation methods and models. Lastly, we propose a mitigation strategy based on ranking across hyperparameters that aims to provide robustness towards such manipulation. This work highlights the difficulty of conducting reliable XAI evaluation and emphasizes the importance of a holistic and transparent approach to evaluation in XAI.

Figures

Figures reproduced from arXiv: 2412.05592 by the authors.

Figure 1
Figure 1. Example of possible faithfulness curves for digit classification. The leftmost curve illustrates how an "intuitive" faithfulness curve might look, while the remaining curves show that there is a lot of variation in how these curves can appear. Illustrating the faithfulness curve Based on the K partitions of S, a set of progres￾sively more perturbed inputs can be created, i.e. {xS1 , · · · , xSK }. Each of the per￾tu… view at source ↗
Figure 2
Figure 2. shows the faithfulness score for each configuration in the feasible set for each dataset. This plot illustrates that the average faithfulness score across the feasible set can often be quite close. However, there is large spread in the scores, which is present for all datasets. This spread demonstrates the lack of robustness in the faithfulness evaluation and is part of the reason why manipu￾lation is possible in th… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. On the Complexity-Faithfulness Trade-off of Gradient-Based Explanations

    cs.LG 2025-08 reject novelty 4.0 of 10

    The paper introduces EF and ΔEF as spectral metrics, but ΔEF is derived from EF, making the complexity-faithfulness trade-off partly tautological.

  2. xai_evals : A Framework for Evaluating Post-Hoc Local Explanation Methods

    cs.LG 2025-02 reject novelty 2.0 of 10

    A technical report introducing xai_evals, a Python package that wraps existing explainability and metric libraries without adding new methods or validated results.

Reference graph

Works this paper leans on

55 extracted references · 35 canonical work pages · cited by 2 Pith papers

  1. [1]

    In: Proceedings of the 32nd International Conference on Neural Information Processing Systems

    Adebayo, J., Gilmer, J., Muelly, M., Goodfellow, I., Hardt, M., Kim, B.: Sanity checks for saliency maps. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. p. 9525–9536. NIPS’18, Curran Associates Inc., Red Hook, NY, USA (2018)

    work page 2018

  2. [2]

    In: Thirty-sixth Conference on Neural Information Processing Sys- tems Datasets and Benchmarks Track (2022),https://openreview.net/forum? id=MU2495w47rz

    Agarwal, C., Krishna, S., Saxena, E., Pawelczyk, M., Johnson, N., Puri, I., Zit- nik, M., Lakkaraju, H.: OpenXAI: Towards a transparent evaluation of model explanations. In: Thirty-sixth Conference on Neural Information Processing Sys- tems Datasets and Benchmarks Track (2022),https://openreview.net/forum? id=MU2495w47rz

    work page 2022

  3. [3]

    In: Advances in Neural Information Processing Sys- tems

    Alvarez Melis, D., Jaakkola, T.: Towards robust interpretability with self- explaining neural networks. In: Advances in Neural Information Processing Sys- tems. pp. – (2018)

    work page 2018

  4. [4]

    In: 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings

    Ancona, M., Ceolini, E., Öztireli, C., Gross, M.: Towards better understanding of gradient-based attribution methods for deep neural networks. In: 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net (2018)

    work page 2018

  5. [5]

    Information Fusion 81, 14–40 (2022)

    Arras, L., Osman, A., Samek, W.: Clevr-xai: A benchmark dataset for the ground truth evaluation of neural network explanations. Information Fusion 81, 14–40 (2022). https://doi.org/https://doi.org/10.1016/j.inffus.2021.11.008 , https://www.sciencedirect.com/science/article/pii/S1566253521002335

    doi:10.1016/j.inffus.2021.11.008 2022

  6. [6]

    CoRR abs/1909.03012 (2019), http://arxiv.org/abs/1909.03012

    Arya, V., Bellamy, R.K.E., Chen, P., Dhurandhar, A., Hind, M., Hoffman, S.C., Houde, S., Liao, Q.V., Luss, R., Mojsilovic, A., Mourad, S., Pedemonte, P., Raghavendra, R., Richards, J.T., Sattigeri, P., Shanmugam, K., Singh, M., Varsh- ney, K.R., Wei, D., Zhang, Y.: One explanation does not fit all: A toolkit and taxonomy of AI explainability techniques. C...

    arXiv 2019

  7. [7]

    PLOS ONE10(7), e0130140 (Jul 2015).https://doi.org/10.1371/ journal.pone.0130140, https://doi.org/10.1371/journal.pone.0130140

    Bach, S., Binder, A., Montavon, G., Klauschen, F., Müller, K.R., Samek, W.: On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PLOS ONE10(7), e0130140 (Jul 2015).https://doi.org/10.1371/ journal.pone.0130140, https://doi.org/10.1371/journal.pone.0130140

    doi:10.1371/journal.pone.0130140 2015

  8. [8]

    In: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR Workshops 2020, Seattle, WA, USA, June 14-19, 2020

    Bansal, N., Agarwal, C., Nguyen, A.: SAM: the sensitivity of attribution methods to hyperparameters. In: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR Workshops 2020, Seattle, WA, USA, June 14-19, 2020. pp. 11–21. Computer Vision Foundation / IEEE (2020)

    work page 2020

Show all 55 references
  1. [9]

    In: Bessiere, C

    Bhatt, U., Weller, A., Moura, J.M.F.: Evaluating and aggregating feature-based model explanations. In: Bessiere, C. (ed.) Proceedings of the Twenty-Ninth Inter- national Joint Conference on Artificial Intelligence, IJCAI 2020. pp. 3016–3022. ijcai.org (2020)

    2020

  2. [10]

    In: International Joint Conference on Artificial Intelligence

    Bhatt, U., Weller, A., Moura, J.M.F.: Evaluating and aggregating feature-based model explanations. In: International Joint Conference on Artificial Intelligence. pp. 3016–3022 (2020).https://doi.org/10.24963/ijcai.2020/417

    2020 doi

  3. [11]

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

    Binder, A., Weber, L., Lapuschkin, S., Montavon, G., Müller, K.R., Samek, W.: Shortcomings of top-down randomization-based sanity checks for evaluations of deep neural network explanations. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (...

    2023

  4. [12]

    Blücher, S., Vielhaben, J., Strodthoff, N.: Decoupling pixel flipping and occlusion strategy for consistent xai benchmarks (2024)

    2024

  5. [13]

    CoRRabs/2203.02928 (2022)

    Brocki, L., Chung, N.C.: Evaluation of interpretability methods and perturbation artifacts in deep neural networks. CoRRabs/2203.02928 (2022)

    2022 arXiv

  6. [14]

    In: Computer Vision – ECCV 2020 Workshops, pp

    Brunke, L., Agrawal, P., George, N.: Evaluating input perturbation methods for interpreting CNNs and saliency map comparison. In: Computer Vision – ECCV 2020 Workshops, pp. 120–134. Springer International Publishing (2020)

    2020

  7. [15]

    Bykov, K., Hedström, A., Nakajima, S., Höhne, M.M.: Noisegrad - enhancing ex- planations by introducing stochasticity to model weights. In: Thirty-Sixth AAAI Conference on Artificial Intelligence, AAAI 2022, Thirty-Fourth Conference on In- novative Applications of Artificial I...

    2022

  8. [16]

    In: III, H.D., Singh, A

    Chalasani, P., Chen, J., Chowdhury, A.R., Wu, X., Jha, S.: Concise explanations of neural networks using adversarial training. In: III, H.D., Singh, A. (eds.) Pro- ceedings of the 37th International Conference on Machine Learning. Proceedings of Machine Learning Research, vol....

    2020

  9. [17]

    In: International Conference on Machine Learning

    Dasgupta, S., Frost, N., Moshkovitz, M.: Framework for evaluating faithfulness of local explanations. In: International Conference on Machine Learning. pp. 4794–

  10. [18]

    In: Computer Vision and Pattern Recognition

    Deng, J., et al.: Imagenet: A large-scale hierarchical image database. In: Computer Vision and Pattern Recognition. pp. 248–255 (2009)

    2009

  11. [19]

    IEEE Signal Processing Magazine29(6), 141–142 (2012)

    Deng, L.: The mnist database of handwritten digit images for machine learning research. IEEE Signal Processing Magazine29(6), 141–142 (2012)

    2012

  12. [20]

    In: IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering, MetroXRAINE 2023, Milano, Italy, October 25-27, 2023

    Dolci, G., Cruciani, F., Galazzo, I.B., Calhoun, V.D., Menegaz, G.: Objective as- sessment of the bias introduced by baseline signals in XAI attribution methods. In: IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering...

    2023

  13. [21]

    Fong, R.C., Vedaldi, A.: Interpretable explanations of black boxes by meaningful perturbation.In:2017IEEEInternationalConferenceonComputerVision(ICCV). pp. 3449–3457 (2017).https://doi.org/10.1109/ICCV.2017.371

    2017 doi

  14. [22]

    In: Ranzato, M., Beygelzimer, A., Dauphin, Y.N., Liang, P., Vaughan, J.W

    Hase, P., Xie, H., Bansal, M.: The out-of-distribution problem in explainability and search methods for feature importance explanations. In: Ranzato, M., Beygelzimer, A., Dauphin, Y.N., Liang, P., Vaughan, J.W. (eds.) Advances in Neural Informa- tion Processing Systems 34: Ann...

    2021

  15. [23]

    In: 2016 CVPR

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: 2016 CVPR. pp. 770–778 (2016).https://doi.org/10.1109/CVPR.2016.90

    2016 doi

  16. [24]

    Transactions on Machine Learning Research (2023), https://openreview.net/forum?id=j3FK00HyfU

    Hedström,A.,Bommer,P.L.,Wickstrøm,K.K.,Samek,W.,Lapuschkin,S.,Höhne, M.M.: The meta-evaluation problem in explainable AI: Identifying reliable esti- mators with metaquantus. Transactions on Machine Learning Research (2023), https://openreview.net/forum?id=j3FK00HyfU

    2023

  17. [25]

    In: Explainable Artificial Intelligence

    Hedström, A., Weber, L., Lapuschkin, S., Höhne, M.: A fresh look at sanity checks for saliency maps. In: Explainable Artificial Intelligence. pp. 403–420. Springer Nature Switzerland, Cham (2024)

    2024

  18. [26]

    Journal of Ma- From Flexibility to Manipulation: The Slippery Slope of XAI Evaluation 17 chine Learning Research24(34), 1–11 (2023),http://jmlr.org/papers/v24/22- 0142.html

    Hedström, A., Weber, L., Krakowczyk, D., Bareeva, D., Motzkus, F., Samek, W., Lapuschkin, S., Höhne, M.M.C.: Quantus: An explainable ai toolkit for re- sponsible evaluation of neural network explanations and beyond. Journal of Ma- From Flexibility to Manipulation: The Slipper...

    2023

  19. [27]

    In: Wallach, H., Larochelle, H., Beygelzimer, A., d 'Alché-Buc, F., Fox, E., Garnett, R

    Hooker, S., Erhan, D., Kindermans, P.J., Kim, B.: A benchmark for in- terpretability methods in deep neural networks. In: Wallach, H., Larochelle, H., Beygelzimer, A., d 'Alché-Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems. vol. 32. Cur...

    2019

  20. [28]

    In: XAI in Action: Past, Present, and Future Applications (2023),https://openreview.net/forum? id=ag1CpSUjPS

    Karimi, A.H., Muandet, K., Kornblith, S., Schölkopf, B., Kim, B.: On the rela- tionship between explanation and prediction: A causal view. In: XAI in Action: Past, Present, and Future Applications (2023),https://openreview.net/forum? id=ag1CpSUjPS

    2023

  21. [29]

    Cell172(5), 1122–1131.e9 (2018).https://doi.org/https: //doi.org/10.1016/j.cell.2018.02.010 , https://www.sciencedirect.com/ science/article/pii/S0092867418301545

    Kermany, D.S., Goldbaum, M., Cai, W., Valentim, C.C., Liang, H., Baxter, S.L., McKeown, A., Yang, G., Wu, X., Yan, F., Dong, J., Prasadha, M.K., Pei, J., Ting, M.Y., Zhu, J., Li, C., Hewett, S., Dong, J., Ziyar, I., Shi, A., Zhang, R., Zheng, L., Hou, R., Shi, W., Fu, X., Duan...

    2018

  22. [30]

    Kindermans, P.J., Hooker, S., Adebayo, J., Alber, M., Schütt, K.T., Dähne, S., Erhan, D., Kim, B.: The (Un)reliability of Saliency Methods, p. 267–280. Springer International Publishing (2019). https://doi.org/10.1007/978-3-030-28954- 6_14

    2019 doi

  23. [31]

    CoRRabs/2404.11330 (2024)

    Koenen, N., Wright, M.N.: Toward understanding the disagreement problem in neural network feature attribution. CoRRabs/2404.11330 (2024)

    2024 arXiv

  24. [32]

    Kokhlikyan, N., Miglani, V., Martin, M., Wang, E., Alsallakh, B., Reynolds, J., Melnikov, A., Kliushkina, N., Araya, C., Yan, S., Reblitz-Richardson, O.: Captum: A unified and generic model interpretability library for pytorch (2020)

    2020

  25. [33]

    CoRR abs/2202.01602 (2022), https://arxiv.org/abs/2202.01602

    Krishna, S., Han, T., Gu, A., Pombra, J., Jabbari, S., Wu, S., Lakkaraju, H.: The disagreement problem in explainable machine learning: A practitioner’s perspec- tive. CoRR abs/2202.01602 (2022), https://arxiv.org/abs/2202.01602

    2022 arXiv

  26. [34]

    Proceedings of the IEEE86(11), 2278–2324 (1998).https: //doi.org/10.1109/5.726791

    Lecun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proceedings of the IEEE86(11), 2278–2324 (1998).https: //doi.org/10.1109/5.726791

    1998 doi

  27. [35]

    In: Proceedings of the 31st International Conference on Neural Information Processing Systems

    Lundberg, S.M., Lee, S.I.: A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. p. 4768–4777. NIPS’17, Curran Associates Inc., Red Hook, NY, USA (2017)

    2017

  28. [36]

    CoRR abs/2208.09473 (2022)

    Mamalakis, A., Barnes, E.A., Ebert-Uphoff, I.: Carefully choose the baseline: Lessons learned from applying XAI attribution methods for regression tasks in geoscience. CoRR abs/2208.09473 (2022)

    2022 arXiv

  29. [37]

    Digital Signal Processing 73, 1–15 (2018)

    Montavon, G., Samek, W., Müller, K.R.: Methods for interpreting and un- derstanding deep neural networks. Digital Signal Processing 73, 1–15 (2018). https://doi.org/https://doi.org/10.1016/j.dsp.2017.10.011 , https: //www.sciencedirect.com/science/article/pii/S1051200417302385

    2018 doi

  30. [38]

    In: Inter- national Conference on Neural Networks

    Morch, N., et al.: Visualization of neural networks using saliency maps. In: Inter- national Conference on Neural Networks. pp. 2085–2090 (1995)

    1995

  31. [39]

    CoRR abs/2007.07584 (2020), https://arxiv.org/abs/2007.07584 18 K

    Nguyen, A., Martínez, M.R.: On quantitative aspects of model interpretability. CoRR abs/2007.07584 (2020), https://arxiv.org/abs/2007.07584 18 K. Wickstrøm et al

    2020 arXiv

  32. [40]

    CoRR abs/2003.08747 (2020), https://arxiv.org/abs/2003.08747

    Rieger, L., Hansen, L.K.: IROF: a low resource evaluation metric for explanation methods. CoRR abs/2003.08747 (2020), https://arxiv.org/abs/2003.08747

    2020 arXiv

  33. [41]

    In: Proceedings of the 39th International Conference on Machine Learning

    Rong, Y., Leemann, T., Borisov, V., Kasneci, G., Kasneci, E.: A consistent and efficient evaluation strategy for attribution methods. In: Proceedings of the 39th International Conference on Machine Learning. pp. 18770–18795. PMLR (2022)

    2022

  34. [42]

    In: International Conference on Machine Learning

    Rong, Y., Leemann, T., Borisov, V., Kasneci, G., Kasneci, E.: A consistent and efficient evaluation strategy for attribution methods. In: International Conference on Machine Learning. pp. 18770–18795 (2022)

    2022

  35. [43]

    IEEE Trans

    Samek, W., Binder, A., Montavon, G., Lapuschkin, S., Müller, K.: Evaluating the visualization of what a deep neural network has learned. IEEE Trans. Neural Networks Learn. Syst.28(11), 2660–2673 (2017)

    2017

  36. [44]

    (eds.): Ex- plainable AI: Interpreting, Explaining and Visualizing Deep Learning

    Samek, W., Montavon, G., Vedaldi, A., Hansen, L.K., Müller, K.R. (eds.): Ex- plainable AI: Interpreting, Explaining and Visualizing Deep Learning. Springer International Publishing (2019).https://doi.org/10.1007/978-3-030-28954-6 , http://dx.doi.org/10.1007/978-3-030-28954-6

    2019 doi

  37. [45]

    In: ICLR Workshop (2015)

    Springenberg, J.T., Dosovitskiy, A., Brox, T., Riedmiller, M.: Striving for simplic- ity: The all convolutional net. In: ICLR Workshop (2015)

    2015

  38. [46]

    Distill (2020)

    Sturmfels, P., Lundberg, S., Lee, S.I.: Visualizing the impact of feature attri- bution baselines. Distill (2020). https://doi.org/10.23915/distill.00022 , https://distill.pub/2020/attribution-baselines

    2020 doi

  39. [47]

    IEEE Transactions on Evolutionary Computation23(5), 828–841 (2019)

    Su, J., Vargas, D.V., Sakurai, K.: One pixel attack for fooling deep neural networks. IEEE Transactions on Evolutionary Computation23(5), 828–841 (2019). https: //doi.org/10.1109/TEVC.2019.2890858

    2019

  40. [48]

    CoRRabs/1806.04205 (2018)

    Sundararajan, M., Taly, A.: A note about: Local explanation methods for deep neu- ral networks lack sensitivity to parameter values. CoRRabs/1806.04205 (2018)

    2018 arXiv

  41. [49]

    In: Precup, D., Teh, Y.W

    Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: Precup, D., Teh, Y.W. (eds.) Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017. Proceedings of Machine Learning Research, v...

    2017

  42. [50]

    In: Precup, D., Teh, Y.W

    Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: Precup, D., Teh, Y.W. (eds.) Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017. Proceedings of Machine Learning Research, v...

    2017

  43. [51]

    CoRR abs/2104.14995 (2021), https://arxiv.org/abs/2104.14995

    Theiner, J., Müller-Budack, E., Ewerth, R.: Interpretable semantic photo geolo- calization. CoRR abs/2104.14995 (2021), https://arxiv.org/abs/2104.14995

    2021 arXiv

  44. [52]

    Tomsett, R., Harborne, D., Chakraborty, S., Gurram, P., Preece, A.D.: Sanity checks for saliency metrics. In: The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, Th...

    2020

  45. [53]

    Wickstrøm, K.K., Trosten, D.J., Løkse, S., Boubekki, A., Mikalsen, K.Ø., Kampffmeyer, M.C., Jenssen, R.: RELAX: representation learning explainability. Int. J. Comput. Vis. pp. 1584–1610 (2023)

    2023

  46. [54]

    Xiao, H., Rasul, K., Vollgraf, R.: Fashion-mnist: a novel image dataset for bench- marking machine learning algorithms (2017)

    2017

  47. [55]

    In: Neural Information Processing Sys- tems (2019)

    Yeh, C.K., Hsieh, C.Y., Suggala, A.S., Inouye, D.I., Ravikumar, P.: On the (in)fidelity and sensitivity of explanations. In: Neural Information Processing Sys- tems (2019)

    2019

Pith tools

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