Pith. sign in

REVIEW 4 major objections 5 minor 35 references

Individualised Counterfactual Examples Using Conformal Prediction Intervals

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read The paper argues that counterfactuals placed where an individual's conformal prediction interval is widest improve that individual's local model accuracy more than distance-only or random counterfactuals.

desk verdict Genuinely novel personalization of counterfactuals via conformal interval width, but the core proxy—wide interval equals informative—is asserted, not established. read the letter →

arxiv 2505.22326 v1 pith:RSLCAJGN submitted 2025-05-28 stat.ML cs.LG

classification stat.MLcs.LG
keywords counterfactualexplanationsconformalpredictionlocallyweighteddataaugmentationintervalwidthindividualisedexplanationbinaryclassificationtransactionfrauddetection
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 argues that when an organisation explains a decision, the single most useful counterfactual to show a person is the one that lies where that person's mental model of the classifier is least certain. It formalises 'least certain' as the width of a conformal prediction interval produced from the person's own small set of known cases, and then selects a counterfactual that flips the predicted class while balancing closeness to the original case against that uncertainty. If this is right, an individual who receives such an explanation builds a more accurate local picture of the decision boundary than someone given a counterfactual chosen purely for proximity, purely for uncertainty, or at random. The experiments confirm the improvement for small training sets on a synthetic hypercube problem and on a realistic transaction-fraud dataset.

What carries the argument

The load-bearing object is the objective function for the CPICF, $$\arg\min_{h_\$\theta$(X) \neq h_\$\theta$(X')} \left( L_{\mathrm{info}}^{($T^{{(k)}}$)}(X') + \$\lambda$ L_{\mathrm{dist}}(X,X') \right),$$ where $L_{\mathrm{info}}^{(T^{(k)})}(X') = 1/C_\alpha(X')$ is the inverse width of the individual's locally weighted conformal prediction interval (a prediction band for the class probability built from locally weighted residuals and a dispersion estimate) and $L_{\mathrm{dist}}$ is a weighted Gower distance for mixed continuous and categorical features. The interval $C_\alpha(X')$ is built from a regression model that the individual would fit to their own subset $T^{(k)}$ of the training data, so the same query can yield different counterfactuals for different individuals. The machinery carries the argument by turning 'informative explanation' into a concrete, computable quantity: a wide interval marks a place where the individual's model is unsure, and the optimisation chooses a nearby opposite-class point in such a place.

What would settle it

Run the paper's single-counterfactual experiment on a dataset with a known true decision boundary, generate counterfactuals at locations of different conformal-interval widths, and compare the local reduction in prediction error; if points from the widest intervals do not reduce the individual's local error more than points from medium-width intervals, the link between interval width and information gain is broken.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that counterfactual explanations can be personalised by explicitly modelling what a particular individual already knows and by using the width of a locally weighted conformal prediction interval as a measure of what the individual would learn from a new point. Minimising $L_{\mathrm{info}}^{(T^{(k)})}(X') = 1/C_\alpha(X')$ together with a weighted Gower distance $\lambda L_{\mathrm{dist}}(X,X')$, under the constraint $h_\theta(X) \neq h_\theta(X')$, produces a conformal prediction interval counterfactual (CPICF). Adding a single CPICF to the individual's training data moves that individual's predicted probabilities closer to the full model's probabilities in a neighbourhood of the original query, and augmenting small training sets with CPICFs improves average precision and F1 score more than unconstrained counterfactual augmentation does.

Load-bearing premise

The load-bearing assumption is that the width of an individual's uncertainty band is a faithful measure of how much a counterfactual placed there will teach that individual; the paper supports this with a thought experiment and indirect evaluations, not with a derivation from the band's coverage guarantee.

Editorial extensions

If this is right

  • A recipient whose local model is improved by the counterfactual will predict the black box's behaviour near their own query more accurately after a single explanation than after a proximity-only or random explanation.
  • For small training sets, data augmentation with CPICFs raises average precision and F1 beyond the unaugmented model and beyond augmentation with unconstrained opposite-class points, which supports using the method when labelled data are scarce.
  • The relative weight $\lambda$ controls the trade-off between usefulness and closeness; values that are too small (uncertainty only) or too large (distance only) fail to improve the individual's model, so the method requires choosing $\lambda$ for the application.
  • Because the counterfactual is computed from the individual's own subset $T^{(k)}$, the same original query can produce different personalised counterfactuals for different individuals, rather than one global explanation.

Reading between the lines

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

  • The authors do not test this, but the same inverse-interval-width score is an active-learning acquisition function: an organisation could use $1/C_\alpha$ to choose which unlabelled points to label for a particular individual, not just which counterfactual to display.
  • A natural extension is to add a diversity or repulsion term to the loss so that a batch of counterfactuals covers several uncertain regions rather than repeating the same one; the paper observes the diversity problem but does not solve it.
  • The observed link between $\alpha$ and the best $\lambda$ suggests a tuning shortcut the authors leave implicit: if interval width is the right informativeness proxy, then $\lambda$ should be set relative to the scale of $1/C_\alpha$, so one could fit a calibration rule mapping coverage error to distance weight instead of searching $\lambda$ afresh for each dataset.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes CPICF, a method for generating individualised counterfactual explanations for a black-box binary classifier. An individual's knowledge is modelled by retraining the same classifier class on a small personal subset T^(k); locally weighted conformal prediction (LWCP) intervals, Eq. (3), are used to define an information score L_info(X') = 1/C_alpha(X'), and the counterfactual is chosen by minimising L_info(X') + lambda * L_dist(X, X') subject to the constraint that the entity's classifier flips its label. The method is demonstrated on a synthetic hypercube dataset and the Tabformer fraud dataset. The experiments evaluate, first, the change in absolute deviation between the individual's predicted probabilities and the entity's oracle probabilities after adding one counterfactual, Eq. (9), and second, data-augmentation gains measured by average precision, F1 score, and ROC AUC. The central claim is that wide conformal prediction intervals identify locations where an additional counterfactual will most improve the individual's local model, and that CPICFs therefore outperform distance-only, uncertainty-only, and unconstrained random counterfactuals.

Significance. If the central heuristic were established, the paper would make a useful contribution to the counterfactual-explanation literature by explicitly modelling the recipient's limited knowledge and connecting counterfactual generation to conformal uncertainty quantification. The paper has clear strengths: it ships code, uses a synthetic setup that permits direct visualisation of decision boundaries and intervals, addresses mixed continuous/categorical tabular data via a weighted Gower distance, and provides falsifiable experimental claims about local probability improvement and augmentation gains. The Tabformer experiments are a plausible real-world stress test. However, the main premise--that interval width is a valid proxy for information gain--is currently an unproven heuristic. The paper's own Conclusion correctly notes that the conformal coverage guarantee is lost once the counterfactual is added, and the experiments require a post hoc, alpha-dependent choice of lambda. These caveats are acknowledged honestly, but they are in tension with the abstract's and conclusion's wording that the approach is validated.

major comments (4)
  1. [Sec. 3.4, Eq. (6), and Eq. (3)] The objective L_info(X') = 1/C_alpha(X') is justified only by the thought experiment in Sec. 3.2, which concerns binary conformal prediction sets at an already-observed point. In the LWCP implementation used in Algorithm 1, the interval width is C_alpha(X') = 2 rho(X') d_alpha, so minimising L_info is equivalent to maximising the estimated mean absolute deviation rho(X') of the individual's model at X'. No analytical argument or direct empirical measurement is provided showing that adding a single labelled counterfactual at a maximum-rho location reduces local predictive error more than at other locations. Wide intervals could reflect irreducible heteroscedastic noise, boundary effects, or regions where one binary label cannot correct a uniformly wrong local model. Because this assumption enters the objective directly, the paper's central claim is currently unsupported.
  2. [Sec. 4.1.1, Fig. 6, and Table 1] The experiments do not isolate the contribution of the width term. In the local-improvement evaluation, the baselines are lambda=0 (pure width), a very large lambda (mostly distance), and unconstrained random counterfactuals; an intermediate lambda is compared, but there is no distance-only counterfactual baseline matched for distance scale, nor a random-augmentation baseline matched for number of added points. In Table 1, CPICF augmentation is compared only to no augmentation, so the observed gains in average precision and F1 score could be produced by adding any reasonable near-boundary examples, not specifically by the conformal-width selection. A direct comparison against random and distance-only augmentation is needed to attribute the improvement to the information term.
  3. [Sec. 4.1.1 and Appendix B] The best value of lambda depends on alpha and is selected post hoc: for alpha=0.1, improvements are reported at lambda=1 or 100, while for alpha=0.2, improvement is reported at lambda=10. The paper acknowledges this and calls for careful calibration of lambda, but the claim that CPICFs improve knowledge is therefore conditional on a tuning parameter that the method does not predict. A principled selection rule, or at least a sensitivity analysis over a wider grid of lambda values with error bars, is necessary before the empirical evidence can be read as supporting the general claim.
  4. [Sec. 5 (Conclusion) and Sec. 3.4] The paper correctly states in the Conclusion that once the counterfactual is added, the conformal prediction guarantees no longer hold because the augmented data are not exchangeable. This is an important limitation: the method's name and justification lean on conformal prediction, but the final procedure has no finite-sample coverage or validity guarantee. The abstract and conclusion nevertheless describe the approach as 'validated' by the experiments. The presentation should clearly label the method as a heuristic whose conformal component provides an uncertainty estimate but no post-hoc guarantee, and the concluding claims should be softened accordingly.
minor comments (5)
  1. [Eq. (4)] In the weighted Gower distance, the categorical sum runs from j=m to p, but if the first m features are continuous, the categorical features are j=m+1,...,p; the index in the second sum should start at m+1.
  2. [Eq. (9)] The expression inside the sum uses Delta_k(X), but the quantity defined in Eq. (7) is Delta_k(X; omega), so the notation should be Delta_k(X; omega) to make the dependence on the perturbation explicit.
  3. [Fig. 6 caption] The caption refers to 'blue, orange and green bars', but the figure shows boxplots rather than bars; additionally, the legend for the side lengths 0.1, 0.5, and 1.0 should be stated in the caption.
  4. [Sec. 4.1.1] The text says 100 sample points are selected without replacement for T^(k) and then 100 CPICFs are generated using sampling with replacement; please clarify whether query points are drawn from T^(k) or from the wider training set, and how the shared starting classifier h_theta_k interacts with the different genetic algorithm seeds.
  5. [Sec. 3.1 and Algorithm 1] The description of the individual's knowledge is ambiguous: Sec. 3.1 says the individual has access to T^(k) and the probabilities or scores of these points, while Algorithm 1 computes the regression targets Y_i = p_theta(X_i) for X_i in T^(k) using the entity's classifier. Please clarify whether the entity computes and reveals these targets or whether the individual already possesses them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the 'most informative counterfactual' claim rests on an explicit heuristic and empirical validation, not on an equation that reduces to its inputs.

full rationale

The paper's proposed objective is an explicit heuristic: L_info(X') = 1/C_alpha(X') is defined (Sec. 3.4) as an inverse conformal interval width, and the claim is that minimizing it subject to class flip selects informative counterfactuals. This is not derived from the evaluation metric Delta(X) in Eq. (9), which measures the change in absolute deviation between the retrained individual model and the oracle after actually adding the counterfactual. No equation equates L_info with Delta, and no parameter fitted to the evaluation data is renamed as a prediction. The scalar lambda is varied over a grid (e.g., lambda in {0,1,100,10^5} in Fig. 6, lambda=1000 in Table 1, with lambda=10 preferred for alpha=0.2 in Appendix B) and the conclusion is explicitly conditional: 'CPICFs can improve the prediction given an appropriate choice of the parameter lambda'. That is hyperparameter selection, not a construction-level circularity. The paper has no load-bearing self-citations: references to conformal prediction and counterfactual methods are external, and no prior work by these authors is used to forbid alternatives. The thought experiment in Sec. 3.2 concerns binary conformal prediction sets and supports the proxy only by analogy; the paper itself flags the absence of theoretical guarantees ('Standard conformal prediction requires the input data to be exchangeable. Thus, once the counterfactual is added, the conformal prediction guarantees no longer hold'). This is an unverified assumption and a correctness/robustness risk, not a case where a 'prediction' reduces by construction to its inputs. I therefore find no significant circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim relies on a small set of domain assumptions about information exchange and on an ad hoc informativeness heuristic; there are no invented entities. The free parameters lambda and alpha are selected empirically and materially affect the conclusions.

free parameters (4)
  • lambda (information-distance trade-off) = 0, 1, 10, 100, 1000, 10^4, 10^5 (varies by dataset and alpha)
    The loss in Eq (6) combines 1/C_alpha and L_dist with weight lambda. The improvement claims hold only for intermediate lambda (e.g., 1, 100 for hypercube alpha=0.1; 10 for alpha=0.2). Lambda is selected empirically per setting.
  • alpha (conformal error level) = 0.1 and 0.2
    Miscoverage level chosen by user; changes the optimal lambda and the qualitative results (Fig 6 vs Fig 8).
  • genetic algorithm hyperparameters = population size 20, 50 evaluations
    Fixed, but not justified; could affect the quality of the counterfactual found by the optimization.
  • size of individual's data subset T^(k) = 100 (hypercube), 4000 (Tabformer dispersion model)
    The individual's knowledge is simulated by sampling T^(k); the size is chosen by the authors and affects interval widths and results.
assumptions (5)
  • domain assumption The individual models the classifier with the same model class as the entity uses (XGBoost).
    Sec 3.1: 'For simplicity we assume that the individual uses the same model class as the original classifier to model their knowledge'.
  • domain assumption The individual has access to the probabilities or scores of points in their subset T^(k) from the black-box classifier.
    Sec 3.1: 'we suppose that an individual k has access to a subset T(k) of the full training dataset T and the probabilities or scores of these points in the full black box classifier', acknowledged as 'a somewhat strong assumption'.
  • standard math Calibration data are exchangeable with future test points, giving the coverage guarantee (2).
    Sec 3.2 uses exchangeability for the split conformal guarantee; Sec 5 notes the guarantee is invalidated once a counterfactual is added.
  • ad hoc to paper A wide conformal prediction interval at X indicates that adding a counterfactual at X reduces the individual's uncertainty most.
    The core heuristic behind Eq (6); introduced via the thought experiment in Sec 3.2 and used to define L_info = 1/C_alpha. No proof is given.
  • ad hoc to paper The change in absolute deviation between oracle and individual probabilities, averaged over a local grid, measures improvement in the individual's knowledge.
    Sec 3.6 defines Delta(X) as this quantity and interprets Delta<0 as improvement; this operationalization is not validated against user studies or other metrics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Individualised Counterfactual Examples Using Conformal Prediction Intervals." pith.science (2026). https://pith.science/paper/RSLCAJGN

@misc{pith2026250522326,
  author       = {Pith},
  title        = {Pith review of: Individualised Counterfactual Examples Using Conformal Prediction Intervals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RSLCAJGN}},
  note         = {Machine review of arXiv:2505.22326}
}
read the original abstract

Counterfactual explanations for black-box models aim to pr ovide insight into an algorithmic decision to its recipient. For a binary classification problem an individual counterfactual details which features might be changed for the model to infer the opposite class. High-dimensional feature spaces that are typical of machine learning classification models admit many possible counterfactual examples to a decision, and so it is important to identify additional criteria to select the most useful counterfactuals. In this paper, we explore the idea that the counterfactuals should be maximally informative when considering the knowledge of a specific individual about the underlying classifier. To quantify this information gain we explicitly model the knowledge of the individual, and assess the uncertainty of predictions which the individual makes by the width of a conformal prediction interval. Regions of feature space where the prediction interval is wide correspond to areas where the confidence in decision making is low, and an additional counterfactual example might be more informative to an individual. To explore and evaluate our individualised conformal prediction interval counterfactuals (CPICFs), first we present a synthetic data set on a hypercube which allows us to fully visualise the decision boundary, conformal intervals via three different methods, and resultant CPICFs. Second, in this synthetic data set we explore the impact of a single CPICF on the knowledge of an individual locally around the original query. Finally, in both our synthetic data set and a complex real world dataset with a combination of continuous and discrete variables, we measure the utility of these counterfactuals via data augmentation, testing the performance on a held out set.

Figures

Figures reproduced from arXiv: 2505.22326 by the authors.

Figure 1
Figure 1. An overview of how the conformal prediction intervals [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. (A) The original classification problem with the points coloured according to their classification (orange and blue), (B) XGB classifier precision-recall curve and average precision (AP) when trained on 60% of the data and tested on a 20% held out fraction, and (C) the decision boundary of the classifier (dashed line). 4. Experiments and discussion Two datasets were investigated to assess the counterfactual generati… view at source ↗
Figure 3
Figure 3. The prediction interval width (darker colour corresponds to smaller prediction interval) for conformal predictors with α = 0.2, based on the same training set, using (A) LWCP or (B) CQR, and (C) size of the conformal prediction set based on direct binary classification The full training data T is overlaid (NB: white re￾gions in (B) correspond to quantile crossing with conflicting (negative) prediction intervals). fe… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Prediction intervals using LWCP for different training datasets, and the same calibration data set. (A1): for the sparse training data in (A2), (B1-2): ablating an area on the left, and (C1-2): ablating all points for x2 > 1 [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: (A) The impact of the parameter λ on the counterfactual selection, showing counterfactual instances for different λ values on the Linfo objective function and (B) CPICFs for different starting instances for the same Ltotal objective function. 12 [PITH_FULL_IMAGE:figur…
Figure 6
Figure 6. Figure 6: A boxplot (left) of the change in the MAD prediction probabilities between with and without the counterfactual in the training set. 100 observations are sampled without replacement in equal proportion from each class in the training data. Each point corresponds to the …
Figure 7
Figure 7. Figure 7: Average precision (area under the precision-recall curve) after training augmented with CPICF examples for the Tabformer dataset, for λ ∈ {0, 10, 104}, α = 0.2, and an unconstrained counterfactual with an alternate label but selected without the distance or conformal l…
Figure 8
Figure 8. Figure 8: The change in the local model prediction following the inclusion of a counterfac￾tual in the training set for 100 repetitions and 7 independent realisations of the hypercube data set, with an error of α = 0.2 While [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages

  1. [1]

    Faithful model explanations through energy-constrained conformal counterfactuals

    Patrick Altmeyer, Mojtaba Farmanbar, Arie van Deursen, and Cynthia CS Liem. Faithful model explanations through energy-constrained conformal counterfactuals. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 10829--10837, 2024

  2. [2]

    Getting a CLUE : A method for explaining uncertainty estimates

    Javier Antoran, Umang Bhatt, Tameem Adel, Adrian Weller, and Jos \'e Miguel Hern \'a ndez-Lobato. Getting a CLUE : A method for explaining uncertainty estimates. In International Conference on Learning Representations, 2021

  3. [3]

    Pymoo: Multi - Objective Optimization in Python

    Julian Blank and Kalyanmoy Deb. Pymoo: Multi - Objective Optimization in Python . IEEE Access, 8: 0 89497--89509, 2020

  4. [4]

    Alex J. Cannon. Non-crossing nonlinear regression quantiles by monotone composite quantile regression neural network, with application to rainfall extremes. Stochastic Environmental Research and Risk Assessment, 32 0 (11): 0 3207--3225, 2018

  5. [5]

    XGBoost : A scalable tree boosting system

    Tianqi Chen and Carlos Guestrin. XGBoost : A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 785--794, 2016

  6. [6]

    Conformal counterfactual inference under hidden confounding

    Zonghao Chen, Ruocheng Guo, Jean-Fran c ois Ton, and Yang Liu. Conformal counterfactual inference under hidden confounding. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 397--408, 2024

  7. [7]

    Beyond pinball loss: Q uantile methods for calibrated uncertainty quantification

    Youngseog Chung, Willie Neiswanger, Ian Char, and Jeff Schneider. Beyond pinball loss: Q uantile methods for calibrated uncertainty quantification. Advances in Neural Information Processing Systems, 34: 0 10971--10984, 2021

  8. [8]

    Learning by transduction

    Alex Gammerman, Volodya Vovk, and Vladimir Vapnik. Learning by transduction. In 14th Conference on Uncertainty in Artificial Intelligence, 1998

Show all 35 references
  1. [9]

    European Union regulations on algorithmic decision-making and a ``right to explanation''

    Bryce Goodman and Seth Flaxman. European Union regulations on algorithmic decision-making and a ``right to explanation''. AI Magazine, 38 0 (3): 0 50--57, 2017

  2. [10]

    Counterfactual explanations and how to find them: literature review and benchmarking

    Riccardo Guidotti. Counterfactual explanations and how to find them: literature review and benchmarking. Data Mining and Knowledge Discovery, 38 0 (5): 0 2770--2824, Sept. 2024

  3. [11]

    Design of experiments for the NIPS 2003 variable selection benchmark

    Isabelle Guyon. Design of experiments for the NIPS 2003 variable selection benchmark. NIPS, 2003

  4. [12]

    Conformal prediction with local weights: randomization enables robust guarantees

    Rohan Hore and Rina Foygel Barber. Conformal prediction with local weights: randomization enables robust guarantees. Journal of the Royal Statistical Society Series B: Statistical Methodology, 87 0 (2): 0 549--578, 2025

  5. [13]

    Santos, and Tom Duckett

    Tomas Krajnik, Joao M. Santos, and Tom Duckett. Life-long spatio-temporal exploration of dynamic environments. In 2015 European Conference on Mobile Robots ( ECMR ) , pages 1--8, Lincoln, United Kingdom, Sept. 2015. IEEE

  6. [14]

    Castle: Cluster-aided space transformation for local explanations

    Valerio La Gatta, Vincenzo Moscato, Marco Postiglione, and Giancarlo Sperl \` . Castle: Cluster-aided space transformation for local explanations. Expert Systems with Applications, 179: 0 115045, 2021

  7. [15]

    A novel augmentation strategy for credit scoring modeling

    Valerio La Gatta, Marco Postiglione, and Giancarlo Sperl \` . A novel augmentation strategy for credit scoring modeling. Neural Computing and Applications, pages 1--13, 2025

  8. [16]

    Tibshirani, and Larry Wasserman

    Jing Lei, Max G’Sell, Alessandro Rinaldo, Ryan J. Tibshirani, and Larry Wasserman. Distribution- Free Predictive Inference for Regression . Journal of the American Statistical Association, 113 0 (523): 0 1094--1111, 2018

  9. [17]

    Conformal inference of counterfactuals and individual treatment effects

    Lihua Lei and Emmanuel J Cand \`e s. Conformal inference of counterfactuals and individual treatment effects. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 83 0 (5): 0 911--938, 2021

  10. [18]

    PUNCC : a Python Library for Predictive Uncertainty Calibration and Conformalization

    Mouhcine Mendil, Luca Mossina, and David Vigouroux. PUNCC : a Python Library for Predictive Uncertainty Calibration and Conformalization . Proceedings of Machine Learning Research, 204: 0 1--20, 2023

  11. [19]

    Efficient classification with counterfactual reasoning and active learning

    Azhar Mohammed, Dang Nguyen, Bao Duong, and Thin Nguyen. Efficient classification with counterfactual reasoning and active learning. In Asian Conference on Intelligent Information and Database Systems, pages 27--38. 2022

  12. [20]

    Improving classification accuracy using data augmentation on small data sets

    Francisco J Moreno-Barea, Jos \'e M Jerez, and Leonardo Franco. Improving classification accuracy using data augmentation on small data sets. Expert Systems with Applications, 161: 0 113696, 2020

  13. [21]

    Explaining Machine Learning classifiers through diverse counterfactual explanations

    Ramaravind Kommiya Mothilal, Amit Sharma, and Chenhao Tan. Explaining Machine Learning classifiers through diverse counterfactual explanations . In Proceedings of the 2020 Conference on Fairness , Accountability , and Transparency , pages 607--617, 2020

  14. [22]

    Tabular transformers for modeling multivariate time series

    Inkit Padhi, Yair Schiff, Igor Melnyk, Mattia Rigotti, Youssef Mroueh, Pierre Dognin, Jerret Ross, Ravi Nair, and Erik Altman. Tabular transformers for modeling multivariate time series. In IEEE International Conference on Acoustics, Speech, and Signal Processing. Institute of...

  15. [23]

    Normalized nonconformity measures for regression Conformal Prediction

    Harris Papadopoulos, Alex Gammerman, and Volodya Vovk. Normalized nonconformity measures for regression Conformal Prediction . AIA '08: Proceedings of 26th IASTED International Conference on Artificial Intelligence and Applications, pages 64--69, 2008

  16. [24]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Ga \"e l Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. Journal of Machine Learning Research, 12: 0 2825--2830, 2011

  17. [25]

    Mocoda: Model-based counterfactual data augmentation

    Silviu Pitis, Elliot Creager, Ajay Mandlekar, and Animesh Garg. Mocoda: Model-based counterfactual data augmentation. Advances in Neural Information Processing Systems, 35: 0 18143--18156, 2022

  18. [26]

    Conformalized quantile regression

    Yaniv Romano, Evan Patterson, and Emmanuel Candes. Conformalized quantile regression. Advances in Neural Information Processing Systems, 32, 2019

  19. [27]

    Learning Relevant Explanations

    Chris Russell, Rory Mc Grath, and Luca Costabello. Learning Relevant Explanations . Retrieved from https://www.researchgate.net , 2020

  20. [28]

    Model-agnostic and scalable counterfactual explanations via reinforcement learning

    Robert-Florian Samoilescu, Arnaud Van Looveren, and Janis Klaise. Model-agnostic and scalable counterfactual explanations via reinforcement learning. arXiv preprint arXiv:2106.02597, 2021

  21. [29]

    Learning optimal conformal classifiers

    David Stutz, Krishnamurthy Dj Dvijotham, Ali Taylan Cemgil, and Arnaud Doucet. Learning optimal conformal classifiers. In International Conference on Learning Representations, 2022

  22. [30]

    Solving the class imbalance problem using a counterfactual method for data augmentation

    Mohammed Temraz and Mark T Keane. Solving the class imbalance problem using a counterfactual method for data augmentation. Machine Learning with Applications, 9: 0 100375, 2022

  23. [31]

    Counterfactual invariance to spurious correlations: why and how to pass stress tests

    Victor Veitch, Alexander D'Amour, Steve Yadlowsky, and Jacob Eisenstein. Counterfactual invariance to spurious correlations: why and how to pass stress tests. In Proceedings of the 35th International Conference on Neural Information Processing Systems, pages 16196--16208, 2021

  24. [32]

    Counterfactual Explanations and Algorithmic Recourses for Machine Learning : A Review

    Sahil Verma, Varich Boonsanong, Minh Hoang, Keegan Hines, John Dickerson, and Chirag Shah. Counterfactual Explanations and Algorithmic Recourses for Machine Learning : A Review . ACM Computing Surveys, 56 0 (12): 0 1--42, Dec. 2024

  25. [33]

    Explainable Artificial Intelligence and Causal Inference Based ATM Fraud Detection

    Yelleti Vivek, Vadlamani Ravi, Abhay Mane, and Laveti Ramesh Naidu. Explainable Artificial Intelligence and Causal Inference Based ATM Fraud Detection . In 2024 IEEE Symposium on Computational Intelligence for Financial Engineering and Economics ( CIFEr ) , pages 1--7, Hoboken...

  26. [34]

    Counterfactual explanations without opening the black box: A utomated decisions and the GDPR

    Sandra Wachter, Brent Mittelstadt, and Chris Russell. Counterfactual explanations without opening the black box: A utomated decisions and the GDPR . Harvard Journal of Law & Technology, 31 0 (2): 0 2018, 2017

  27. [35]

    A comprehensive survey on data augmentation

    Zaitian Wang, Pengfei Wang, Kunpeng Liu, Pengyang Wang, Yanjie Fu, Chang-Tien Lu, Charu C Aggarwal, Jian Pei, and Yuanchun Zhou. A comprehensive survey on data augmentation. arXiv preprint arXiv:2405.09591, 2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.