Pith. sign in

REVIEW 3 major objections 6 minor 40 references

A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper argues that one objective with a single trade-off parameter λ contains four existing counterfactual explanation approaches as special cases or limits while remaining cheaper to solve than causal recourse.

desk verdict A clean but incrementally novel weighted counterfactual objective, undermined by an overstated theorem and anecdotal experiments. read the letter →

arxiv 2505.02435 v2 pith:5IILLDKT submitted 2025-05-05 cs.LG cs.AIstat.ML

Pouria Fatemi , Ehsan Sharifian , Mohammad Hossein Yassaee This is my paper · ORCID
classification cs.LGcs.AIstat.ML
keywords counterfactualexplanationsbacktrackingcounterfactualscausalalgorithmicrecoursestructuralmodelsbijectivegenerationmechanismmodelinterpretabilityactionableParetooptimality
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 proposes a single optimization objective for generating counterfactual explanations that are both close to the observed input and consistent with a known causal model. The objective minimizes the sum of a distance in input space and a weighted distance in the latent-noise space, subject to the counterfactual flipping the model's prediction. The paper argues that this objective, with one trade-off parameter $\lambda$, contains four existing counterfactual approaches as special cases or limits: classical counterfactual explanations ($\lambda=0$), deep backtracking explanations ($\lambda\to\infty$), backtracking counterfactual explanations under a specific backtracking distribution, and causal algorithmic recourse under convexity and linearity assumptions. If true, this gives a computationally cheaper way to produce causally grounded, actionable explanations without the combinatorial subset search or latent-distribution integration of prior methods.

What carries the argument

The load-bearing object is the invertible reduced-form map $X=F(U)$ between independent latent noises and observed variables, which makes $U=F^{-1}(X)$ deterministic and lets the method treat a counterfactual as a nearby latent vector rather than a modified causal graph. On top of this sits the BRACE objective with two distances: observed-space fidelity $d_X$ and latent-space plausibility $d_U$, balanced by $\lambda$. The proof machinery is Theorem 5.1's identity $u^{ICF}_A=F^{-1}(F_A(F^{-1}(x)))$, which rewrites interventional counterfactuals as backtracking counterfactuals, and the Pareto scalarization argument used in Theorem 6.1 to show that varying $\lambda$ can reach the point where a constrained latent deviation is traded optimally.

What would settle it

Take a linear SCM with $F$ and $h$ linear, enumerate the full Pareto frontier of the vector optimization (19), and check whether for every feasible latent deviation $\alpha$ the scalarized objective (8) reaches the constrained optimum of (17) for some $\lambda$; any missed point falsifies the existence claim in Theorem 6.1. A simpler check: choose a non-injective $F$ such as $X_1=U_1$, $X_2=U_1^2+U_2$; then $F^{-1}(x)$ is undefined for some $x$, and BRACE cannot be run as written.

Watch

Extended reading notes

Core claim

Under a known structural causal model with invertible reduced form $X=F(U)$, BRACE solves $$\min_{$x^{{CF}}$,$u^{{CF}}$} d_X(x,$x^{{CF}}$)+\$\lambda$ d_U(u,$u^{{CF}}$) \quad \text{s.t.}\quad h($x^{{CF}}$)=$y^{{CF}}$,\ $x^{{CF}}$=F($u^{{CF}}$),\ x=F(u).$$ The paper's central claim is that this scalarized two-distance objective unifies four existing paradigms: at $\lambda=0$ it reduces to classical counterfactual explanations, as $\lambda\to\infty$ it reduces to deep backtracking explanations, with the backtracking kernel $P_B(u^{CF}\mid u)\propto\exp(-d_X(F(u),F(u^{CF}))-\lambda d_U(u,u^{CF}))$ it coincides with backtracking counterfactual explanations, and under convex distances and linear $F,h$ it yields counterfactuals at least as close to $x$ as causal algorithmic recourse for the same latent deviation. The paper further establishes Theorem 5.1, that under the bijective generation mechanism backtracking counterfactuals contain interventional counterfactuals as a special case, via the construction $u^{ICF}_A=F^{-1}(x^{ICF})$.

Load-bearing premise

The whole construction assumes a known structural causal model with independent noise and no hidden confounders, whose reduced-form map $F$ is invertible; if $F^{-1}$ is not well-defined or the causal graph is misspecified, the latent vector $u$ cannot be recovered and the proofs of Theorem 5.1 and the equivalence to backtracking explanations collapse.

Editorial extensions

If this is right

  • Setting $\lambda=0$ makes BRACE exactly the classical counterfactual explanation objective, and letting $\lambda\to\infty$ makes it exactly deep backtracking explanations; the same implementation covers both extremes.
  • BRACE is equivalent to backtracking counterfactual explanations with the exponential kernel in Eq. (14), so it gives a tractable route to backtracking counterfactuals that avoids marginalizing over the backtracking distribution.
  • Under convex distances and linear $F$ and $h$, for every fixed latent deviation $\alpha$ there is a $\lambda$ whose BRACE counterfactual is no farther from the observed input than causal algorithmic recourse's counterfactual; because BRACE avoids the exponential subset search, it is also cheaper to compute.
  • The identity $u^{ICF}_A=F^{-1}(F_A(F^{-1}(x)))$ implies that any interventional counterfactual can be realized as a backtracking counterfactual once the model is invertible.
  • In the loan-approval simulations, BRACE changes both the loan amount and the repayment duration, keeping the implied monthly payment closer to the original than the baselines, which is the sense in which the explanations are actionable.

Reading between the lines

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

  • Editorial extension: the theory only needs invertibility of $F$, not linearity, so BRACE should be testable on nonlinear invertible generative models such as normalizing flows; if the dominance over causal algorithmic recourse persists there, the method extends well beyond the paper's linear experiments.
  • The exponential backtracking kernel in Eq. (14) is a substantive modeling commitment, not a neutral choice; a user whose true backtracking distribution differs would obtain different counterfactuals from the same objective.
  • The parameter $\lambda$ can be read as a Lagrange multiplier on latent deviation, which suggests selecting it empirically by downstream actionability measures (such as effort or cost of the recommended change) rather than by default values.
  • Theorem 6.1's argument implies a general comparison recipe: any recourse method that fixes a latent-space deviation can be dominated by re-optimizing observed-space distance inside that deviation; a direct test would compare BRACE with causal algorithmic recourse on a nonlinear SCM using a sweep over $\lambda$.
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 / 6 minor

Summary. The manuscript proposes BRACE, a counterfactual explanation method that solves the scalarized objective in Eq. (8): minimize dX(x, xCF) + λ dU(u, uCF) subject to h(xCF) = yCF and xCF = F(uCF), assuming an invertible structural causal model X = F(U). The authors claim that this objective unifies classical Counterfactual Explanations (λ = 0), Deep Backtracking Explanations (λ → ∞), Backtracking Counterfactual Explanations via the Gibbs-type PB in Eq. (14), and Causal Algorithmic Recourse under convexity and linearity assumptions (Theorem 6.1). The paper also proves, under the bijective generation mechanism, that interventional counterfactuals are a special case of backtracking counterfactuals (Theorem 5.1), and evaluates BRACE on a small simulation based on the German Credit dataset.

Significance. The BRACE objective is a natural and computationally attractive formulation, and the reduction of Causal Algorithmic Recourse's combinatorial search to a continuous optimization is a legitimate potential advantage. The derivations in Appendix B and Theorem 5.1 are largely correct: the former shows that BRACE can be recovered as a backtracking counterfactual for a particular choice of PB, and the latter establishes that interventional counterfactuals are among the backtracking solutions when F is invertible. However, the claimed theoretical superiority over Causal Algorithmic Recourse in Theorem 6.1 is false as stated; the counterexample in Major Comment 1 is within the theorem's hypotheses. Since this theorem is one of the paper's advertised unifications, the claim needs substantial repair. If the theorem is corrected to a weaker 'no worse' statement or with additional assumptions on the intervention set, the remaining framework may still be valuable.

major comments (3)
  1. [Section 6, Theorem 6.1] The theorem is false as stated. Take the identity SCM X = U (i.e., F is the identity), h(X) = X1 - X2, observed x = (0,0), target yCF = 1, and dX = dU = ℓ1. In CAR (Eq. (5)) with cost(A; x) = dX(xCF, x), intervening on the full set A = {1,2} gives xCF = (1,0), which has dX = 1 and satisfies h(xCF) = 1; this is the unconstrained minimum of dX on the constraint set. For BRACE (Eq. (8)), the objective becomes (1+λ)(|x1|+|x2|) subject to x1 - x2 = 1, and its minimum has dX = 1 for every λ ≥ 0. Therefore no λ yields a BRACE counterfactual with dX strictly smaller than CAR's optimum, contradicting the theorem's claim that BRACE 'outperforms' CAR. The root cause is that Theorem 6.1 imposes no restriction on the intervention set A, so CAR can always choose the full set and match the global CE minimum; a correct theorem would need to exclude full-set interventions, add an action cost, or claim only 'no worse'.
  2. [Section 6, proof of Theorem 6.1] The proof does not establish the conclusion even in a weaker form. Inequality (18) is dX(x*BCF, x) ≤ dX(x*ICF, x), which is non-strict; the statement that 'other feasible values ... may reduce the objective further' does not prove strict improvement. In addition, the assertion that scalarization (20) with λ > 0 yields all Pareto-optimal points of the vector optimization (19) is not generally true: for a non-strictly convex Pareto frontier, points can have supporting normals with zero dU weight and thus are not minima of dX + λ dU for any λ > 0. The argument also does not show that, for the specific α in (16), there exists a λ whose solution of (8) satisfies dU(uCF, u) = α. Convexity and linearity alone do not guarantee this. The proof therefore needs a different argument if the theorem is to be salvaged.
  3. [Section 9 and Abstract] The paper's conclusion and abstract state that BRACE 'generalizes' Causal Algorithmic Recourse and yields counterfactuals closer to the original input than CAR, but the only theoretical support for this claim is Theorem 6.1, which is false as stated. The experiments in Section 8 do not directly compare the dX distances achieved by BRACE and CAR; Table 1 reports feature values and monthly payments, but not the distance to the original input. The conclusion and the paper's claims of outperformance should be rewritten to reflect whatever corrected version of Theorem 6.1 is adopted.
minor comments (6)
  1. [Section 2.1, Eq. (3)] The paper assumes a Bijective Generation Mechanism, which is strong; the authors should state more explicitly that all subsequent results (Theorem 5.1, Appendix B, and the CAR comparison) depend on this invertibility and that the method may not apply to non-invertible SCMs.
  2. [Section 5, after Theorem 5.1] The claim that Theorem 5.1 is 'the first result that relates backtracking and interventional counterfactuals' is misleading, since von Kugelgen et al. (2023) already discuss the relationship between these paradigms; the novelty statement should be softened.
  3. [Section 8] The method is repeatedly claimed to be computationally more efficient than Causal Algorithmic Recourse, but no runtime or complexity measurements are reported; a comparison of wall-clock times would substantiate the claim.
  4. [Section 8.2, Eq. (25)] The optimization problem uses normalized features (dividing by σi), but Table 1 reports values in original units; the relationship between the objective in (25) and the reported counterfactuals should be clarified.
  5. [Eq. (14)] The backtracking conditional distribution is written without its normalizing constant; the authors should note that it is defined up to proportionality.
  6. [Proof of Theorem 5.1] The proof does not explicitly justify that uICF_A equals u on indices not in A; this follows from the recursive inversion of the structural equations and should be stated to make the proof self-contained.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the λ-limit unifications are explicit special cases and the PB choice is a construction, not a smuggled assumption.

full rationale

The derivation chain is self-contained and not circular. BRACE's objective (8) is defined as the sum of an observed-space distance and a latent-space distance; therefore the λ=0 and λ→∞ limits literally reproduce the CE objective (4) and the DBE objective (7). This is an explicit parameter-limit relationship, not a fitted prediction and not a hidden definition of the target result: the paper does not use CE or DBE as evidence for BRACE, it simply notes that they are special cases of a newly proposed objective. Similarly, Eq. (14) is a deliberately chosen backtracking distribution whose Gibbs form has the BRACE objective as its negative exponent; Appendix B shows the MAP estimate of the resulting BCF problem is exactly BRACE. That is a construction/embedding, not a circular derivation. Theorem 5.1 constructs uICF_A = F^{-1}(xICF) and substitutes it into the backtracking equations, a direct consequence of invertibility; it is a proof step about the relation between two counterfactual definitions, not a circularity. The CAR comparison (Theorem 6.1) is an attempted external comparison; its strict 'closer' conclusion is not fully supported by the non-strict inequality (18), but that is a correctness/mathematical-strength concern, not a self-referential reduction. No fitted parameter is renamed as a prediction, and no load-bearing claim is justified by a self-citation chain. The paper's own Section 10 limitations (convexity requirement, simple test models) are acknowledged scope restrictions, not circular inputs.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities; BRACE is an objective function, and the backtracking distribution in Eq (14) is chosen to match the objective rather than being an independent postulate.

free parameters (2)
  • trade-off parameter lambda = 1 and 1.2 in experiments
    Controls the balance between input-space distance and latent-space distance in Eq (8); the paper picks lambda=1 and lambda=1.2 by hand, and the resulting counterfactuals and the claimed actionability change with lambda.
  • constraint penalty beta = not specified
    In Eq (23) the output constraint is imposed with an increasing penalty beta, but the schedule used in the experiments is not given.
assumptions (6)
  • domain assumption Bijective Generation Mechanism: F is invertible
    Stated in Section 2.1; needed to compute u=F^{-1}(x) and to prove Theorem 5.1 and Appendix B.
  • domain assumption Causal sufficiency: no hidden confounders
    Stated in Section 2.1; ensures the noise variables U fully determine X through F.
  • domain assumption Known structural equations S and known causal graph
    Assumed in Section 2.3; the method requires the full SCM to evaluate F and F^{-1}.
  • domain assumption Mutual independence of U components
    Assumed in Section 2.1; the paper notes in Section 10 that the counterfactual UCF may not preserve independence.
  • ad hoc to paper For Theorem 6.1: dX and dU convex, F and h linear, cost(A;x)=dX(xCF,x)
    Introduced in Theorem 6.1 to guarantee that all Pareto-optimal points of the vector optimization are reachable by scalarization.
  • domain assumption Distance functions chosen as l1 for dX and l2 for dU
    Section 7 selects these metrics for interpretability and smoothness; the theoretical results are stated for general convex distances.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability." pith.science (2026). https://pith.science/paper/5IILLDKT

@misc{pith2026250502435,
  author       = {Pith},
  title        = {Pith review of: A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5IILLDKT}},
  note         = {Machine review of arXiv:2505.02435}
}
read the original abstract

Counterfactual explanations enhance interpretability by identifying alternative inputs that produce different outputs, offering localized insights into model decisions. However, traditional methods often neglect causal relationships, leading to unrealistic examples. While newer approaches integrate causality, they are computationally expensive. To address these challenges, we propose an efficient method called BRACE based on backtracking counterfactuals that incorporates causal reasoning to generate actionable explanations. We first examine the limitations of existing methods and then introduce our novel approach and its features. We also explore the relationship between our method and previous techniques, demonstrating that it generalizes them in specific scenarios. Finally, experiments show that our method provides deeper insights into model outputs.

Figures

Figures reproduced from arXiv: 2505.02435 by the authors.

Figure 1
Figure 1. Causal graph of the bank’s high-risk detection model. X1 is gender, X2 is age, X3 is loan amount, and X4 is repayment duration in months. The model’s output Yˆ indicates high or low risk for loan approval. function as penalty terms: arg min xCF dX(x, x CF) + λ dU (F −1 (x), F −1 (x CF)) + β Loss h(x CF), yCF , (23) where Loss(.) is a common classification loss, such as cross-entropy. To approximate solutions in pra… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 28 canonical work pages

  1. [1]

    write newline

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

  2. [2]

    J., and Coskun, A

    Ates, E., Aksar, B., Leung, V. J., and Coskun, A. K. Counterfactual explanations for multivariate time series. In 2021 International Conference on Applied Artificial Intelligence (ICAPAI), pp.\ 1--8. IEEE, 2021

    work page 2021

  3. [3]

    L., Hsu, W., and Tan, N

    Bhoi, S., Lee, M. L., Hsu, W., and Tan, N. C. Refine: a fine-grained medication recommendation system using deep learning and personalized drug interaction modeling. Advances in Neural Information Processing Systems, 36, 2024

    work page 2024

  4. [4]

    Boyd, S. P. and Vandenberghe, L. Convex optimization. Cambridge university press, 2004

    2004

  5. [5]

    Explanations based on the missing: Towards contrastive explanations with pertinent negatives

    Dhurandhar, A., Chen, P.-Y., Luss, R., Tu, C.-C., Ting, P., Shanmugam, K., and Das, P. Explanations based on the missing: Towards contrastive explanations with pertinent negatives. Advances in neural information processing systems, 31, 2018

    work page 2018

  6. [6]

    H., and Sch \"o lkopf, B

    Dominguez-Olmedo, R., Karimi, A. H., and Sch \"o lkopf, B. On the adversarial robustness of causal algorithmic recourse. In International Conference on Machine Learning, pp.\ 5324--5342. PMLR, 2022

    work page 2022

  7. [7]

    Against counterfactual miracles

    Dorr, C. Against counterfactual miracles. The Philosophical Review, 125 0 (2): 0 241--286, 2016

    work page 2016

  8. [8]

    and Kim, B

    Doshi-Velez, F. and Kim, B. Towards a rigorous science of interpretable machine learning. arXiv preprint arXiv:1702.08608, 2017

    arXiv 2017

Show all 40 references
  1. [9]

    Asymmetric shapley values: incorporating causal knowledge into model-agnostic explainability

    Frye, C., Rowat, C., and Feige, I. Asymmetric shapley values: incorporating causal knowledge into model-agnostic explainability. Advances in Neural Information Processing Systems, 33: 0 1229--1239, 2020

    2020

  2. [10]

    G., and Claassen, T

    Heskes, T., Sijben, E., Bucur, I. G., and Claassen, T. Causal shapley values: Exploiting causal knowledge to explain individual predictions of complex models. Advances in neural information processing systems, 33: 0 4778--4789, 2020

    2020

  3. [11]

    Statlog (German Credit Data)

    Hofmann, H. Statlog (German Credit Data) . UCI Machine Learning Repository, 1994. DOI : https://doi.org/10.24432/C5NC77

    1994 doi

  4. [12]

    Feature relevance quantification in explainable ai: A causal problem

    Janzing, D., Minorics, L., and Bl \"o baum, P. Feature relevance quantification in explainable ai: A causal problem. In International Conference on artificial intelligence and statistics, pp.\ 2907--2916. PMLR, 2020

    2020

  5. [13]

    Have we learned to explain?: How interpretability methods can learn to encode predictions in their interpretations

    Jethani, N., Sudarshan, M., Aphinyanaphongs, Y., and Ranganath, R. Have we learned to explain?: How interpretability methods can learn to encode predictions in their interpretations. In International Conference on Artificial Intelligence and Statistics, pp.\ 1459--1467. PMLR, 2021

    2021

  6. [14]

    On measuring causal contributions via do-interventions

    Jung, Y., Kasiviswanathan, S., Tian, J., Janzing, D., Bl \"o baum, P., and Bareinboim, E. On measuring causal contributions via do-interventions. In International Conference on Machine Learning, pp.\ 10476--10501. PMLR, 2022

    2022

  7. [15]

    Model-agnostic counterfactual explanations for consequential decisions

    Karimi, A.-H., Barthe, G., Balle, B., and Valera, I. Model-agnostic counterfactual explanations for consequential decisions. In International conference on artificial intelligence and statistics, pp.\ 895--905. PMLR, 2020 a

    2020

  8. [16]

    u gelgen, J., Sch \

    Karimi, A.-H., Von K \"u gelgen, J., Sch \"o lkopf, B., and Valera, I. Algorithmic recourse under imperfect causal knowledge: a probabilistic approach. Advances in neural information processing systems, 33: 0 265--277, 2020 b

    2020

  9. [17]

    Algorithmic recourse: from counterfactual explanations to interventions

    Karimi, A.-H., Sch \"o lkopf, B., and Valera, I. Algorithmic recourse: from counterfactual explanations to interventions. In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp.\ 353--362, 2021

    2021

  10. [18]

    A survey of algorithmic recourse: contrastive explanations and consequential recommendations

    Karimi, A.-H., Barthe, G., Sch \"o lkopf, B., and Valera, I. A survey of algorithmic recourse: contrastive explanations and consequential recommendations. ACM Computing Surveys, 55 0 (5): 0 1--29, 2022

    2022

  11. [19]

    On the relationship between explanation and prediction: A causal view

    Karimi, A.-H., Muandet, K., Kornblith, S., Sch \"o lkopf, B., and Kim, B. On the relationship between explanation and prediction: A causal view. In XAI in Action: Past, Present, and Future Applications, 2023. URL https://openreview.net/forum?id=ag1CpSUjPS

    2023

  12. [20]

    L., Julian, K., and Kochenderfer, M

    Katz, G., Barrett, C., Dill, D. L., Julian, K., and Kochenderfer, M. J. Reluplex: An efficient smt solver for verifying deep neural networks. In Computer Aided Verification: 29th International Conference, CAV 2017, Heidelberg, Germany, July 24-28, 2017, Proceedings, Part I 30,...

    2017

  13. [21]

    Kim, B., Khanna, R., and Koyejo, O. O. Examples are not enough, learn to criticize! criticism for interpretability. Advances in neural information processing systems, 29, 2016

    2016

  14. [22]

    u gelgen, J., Sch \

    Kladny, K.-R., von K \"u gelgen, J., Sch \"o lkopf, B., and Muehlebach, M. Deep backtracking counterfactuals for causally compliant explanations. Transactions on Machine Learning Research, 2024

    2024

  15. [23]

    Inverse classification for comparison-based interpretability in machine learning

    Laugel, T., Lesot, M.-J., Marsala, C., Renard, X., and Detyniecki, M. Inverse classification for comparison-based interpretability in machine learning. arXiv preprint arXiv:1712.08443, 2017

    2017 arXiv

  16. [24]

    Lundberg, S. M. and Lee, S.-I. A unified approach to interpreting model predictions. Advances in neural information processing systems, 30, 2017

    2017

  17. [25]

    Explanation in artificial intelligence: Insights from the social sciences

    Miller, T. Explanation in artificial intelligence: Insights from the social sciences. Artificial intelligence, 267: 0 1--38, 2019

    2019

  18. [26]

    Interpretable machine learning

    Molnar, C. Interpretable machine learning. Lulu. com, 2020

    2020

  19. [27]

    Counterfactual identifiability of bijective causal models

    Nasr-Esfahany, A., Alizadeh, M., and Shah, D. Counterfactual identifiability of bijective causal models. In International Conference on Machine Learning, pp.\ 25733--25754. PMLR, 2023

    2023

  20. [28]

    Causality

    Pearl, J. Causality. Cambridge university press, 2009

    2009

  21. [29]

    Elements of causal inference: foundations and learning algorithms

    Peters, J., Janzing, D., and Sch \"o lkopf, B. Elements of causal inference: foundations and learning algorithms. MIT press, 2017

    2017

  22. [30]

    Face: feasible and actionable counterfactual explanations

    Poyiadzi, R., Sokol, K., Santos-Rodriguez, R., De Bie, T., and Flach, P. Face: feasible and actionable counterfactual explanations. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, pp.\ 344--350, 2020

    2020

  23. [31]

    why should i trust you?

    Ribeiro, M. T., Singh, S., and Guestrin, C. " why should i trust you?" explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, pp.\ 1135--1144, 2016

    2016

  24. [32]

    Curious exploration via structured world models yields zero-shot object manipulation

    Sancaktar, C., Blaes, S., and Martius, G. Curious exploration via structured world models yields zero-shot object manipulation. Advances in Neural Information Processing Systems, 35: 0 24170--24183, 2022

    2022

  25. [33]

    Counterfactual explanations can be manipulated

    Slack, D., Hilgard, A., Lakkaraju, H., and Singh, S. Counterfactual explanations can be manipulated. Advances in neural information processing systems, 34: 0 62--75, 2021

    2021

  26. [34]

    Intriguing properties of neural networks

    Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I., and Fergus, R. Intriguing properties of neural networks. arXiv preprint arXiv:1312.6199, 2013

    2013 arXiv

  27. [35]

    u gelgen, J., Karimi, A.-H., Bhatt, U., Valera, I., Weller, A., and Sch \

    Von K \"u gelgen, J., Karimi, A.-H., Bhatt, U., Valera, I., Weller, A., and Sch \"o lkopf, B. On the fairness of causal algorithmic recourse. In Proceedings of the AAAI conference on artificial intelligence, volume 36, pp.\ 9584--9594, 2022

    2022

  28. [36]

    Backtracking counterfactuals

    Von K \"u gelgen, J., Mohamed, A., and Beckers, S. Backtracking counterfactuals. In Conference on Causal Learning and Reasoning, pp.\ 177--196. PMLR, 2023

    2023

  29. [37]

    Counterfactual explanations without opening the black box: Automated decisions and the gdpr

    Wachter, S., Mittelstadt, B., and Russell, C. Counterfactual explanations without opening the black box: Automated decisions and the gdpr. Harv. JL & Tech., 31: 0 841, 2017

    2017

  30. [38]

    Shapley flow: A graph-based approach to interpreting model predictions

    Wang, J., Wiens, J., and Lundberg, S. Shapley flow: A graph-based approach to interpreting model predictions. In International Conference on Artificial Intelligence and Statistics, pp.\ 721--729. PMLR, 2021

    2021

  31. [39]

    Pixiu: A comprehensive benchmark, instruction dataset and large language model for finance

    Xie, Q., Han, W., Zhang, X., Lai, Y., Peng, M., Lopez-Lira, A., and Huang, J. Pixiu: A comprehensive benchmark, instruction dataset and large language model for finance. Advances in Neural Information Processing Systems, 36, 2024

    2024

  32. [40]

    and Bareinboim, E

    Zhang, J. and Bareinboim, E. Fairness in decision-making—the causal explanation formula. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32, 2018

    2018

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.