New measure SEV shows non-sparse models can still have low decision sparsity.
problem Non-sparse models can still make accurate decisions based on a few features.
method Introduced Sparse Explanation Value (SEV) to measure decision sparsity, not overall model sparsity.
result Many non-sparse models have low decision sparsity, as measured by SEV.
New research challenges the idea that counterfactual explanations should be sparse.
problem Predictive multiplicity leads to multiple models giving almost equal solutions.
method Derive a general upper bound for counterfactual costs under multiplicity and compare sparse vs. data support approaches.
result Data support methods are more robust to multiplicity but have higher counterfactual costs.
SpArX creates faithful explanations of neural networks' decision-making.
problem Challenges in explaining neural networks' decisions.
method Sparsifies MLPs while maintaining structure, then translates into QAFs for argumentative explanations.
result SpArX provides more faithful explanations than existing methods.
Method trains deep models to explain predictions with fewer examples.
problem Difficulty in humans understanding deep model predictions.
method Simultaneously trains prediction and explanation models with sparse regularization.
result Improves faithfulness of explanations with fewer examples while maintaining predictive performance.
New method identifies differences between groups in low-dimensional data representations.
problem Identifying meaningful differences between groups in low-dimensional data representations.
method Introduce Global Counterfactual Explanation (GCE) and Transitive Global Translations (TGT) for computing GCEs.
result TGT identifies sparse, accurate explanations that match real data patterns.
Develops transparent global models consistent with local explanations.
problem Creating globally interpretable models that align with local explanations from black-box models.
method Custom boolean features from sparse local contrastive explanations are used to train a globally transparent model.
result Custom transparent models have higher local consistency compared to other strategies.
Generative models explain machine learning predictions with counterfactual instances.
problem Generating human-interpretable insights into machine learning model predictions.
method Sparse counterfactual explanations using conditional generative models.
result Single forward pass generates batches of counterfactual instances.
New methods create counterfactuals for image regression models.
problem Creating interpretable explanations for regression models in images.
method Two methods using diffusion-based generative models to create counterfactuals.
result Diffusion-based methods produce realistic, semantic, and smooth counterfactuals.
TRUST improves tree models' accuracy while maintaining interpretability.
problem Piecewise-constant regression trees lack in predictive accuracy compared to black-box models.
method Combines Random Forest accuracy with interpretability of shallow trees and sparsity of linear models, using LLMs for explanations.
result TRUST outperforms other interpretable models in predictive accuracy and matches Random Forest's accuracy.
Gradient-based explanations correlate with Android malware classifier robustness.
problem Evasion attacks on Android malware classifiers using sparse perturbations.
method Investigated gradient-based attribution methods for explaining classifier decisions and their evenness, proposing metrics to assess adversarial robustness.
result Gradient-based explanations, especially Integrated Gradients, correlate with adversarial robustness of malware classifiers.
We show new connections between adversarial learning and explainability for deep neural networks (DNNs). One form of explanation of the output of a neural network model in terms of its input features, is a vector of feature-attributions. Two desirable characteristics of an attribution-based explanation are: (1) $\texti…
Proposes sparse local and regional counterfactual rules for robust recourses.
problem Challenges in counterfactual explanations, especially stability, synthesis, and implementation.
method Probabilistic framework using Random Forest to derive sparse local and regional counterfactual rules.
result Effective recourses derived from high-density regions, providing sparse and robust counterfactual rules.
Sparse linear models improve neural network debuggability.
problem Improving neural network interpretability and debugging.
method Using sparse linear models over learned deep feature representations.
result The approach leads to more debuggable and accurate neural networks.
Proposes a method for multilevel explanations of black-box models.
problem Need for explanations at intermediate or group levels, especially for GDPR compliance.
method Meta-method that builds a multilevel explanation tree using local explainability methods.
result Effective multilevel explanations for groups of data points, including novel test points.
Paper introduces Native Guide for generating time series counterfactual explanations.
problem Lack of explainability for time series data in AI systems.
method Model-agnostic, instance-based counterfactual generation for time series classification.
result Native Guide produces better counterfactual explanations than benchmarks.
Multi-SpaCE generates valid counterfactual explanations for multivariate time series data.
problem Lack of transparency in deep learning models for multivariate time series data.
method Multi-objective counterfactual explanation method using NSGA-II for multivariate time series data.
result Ensures perfect validity and superior performance compared to existing methods.
The paper proposes a method to generate diverse counterfactual explanations for anomaly detection in time series data.
problem Lack of helpful explanations for anomaly detection models in time series data.
method Model-agnostic algorithm that generates diverse counterfactual examples for anomaly detection models.
result The method produces counterfactual examples that are not considered anomalous by the detection model and satisfy validity, plausibility, and closeness criteria.
Adaptive regularization prevents overfitting in large-scale sparse feature models.
problem Overfitting in models with large-scale sparse categorical features.
method Adaptive regularization of embedding layers' norm budget.
result Improves model performance within a single epoch and prevents multi-epoch performance degradation.
SISR improves feature attribution in complex payoff schemes.
problem Distorted feature attributions due to non-additive payoff functions and high-dimensional feature spaces.
method Sparse Isotonic Shapley Regression (SISR) learns a monotonic transformation to restore additivity and enforces L0 sparsity.
result SISR achieves strong support recovery and stable attributions across various payoff schemes.
Sparse principal component analysis (sparse PCA) aims at finding a sparse basis to improve the interpretability over the dense basis of PCA, meanwhile the sparse basis should cover the data subspace as much as possible. In contrast to most of existing work which deal with the problem by adding some sparsity penalties o…
New method discovers concepts in hidden feature layers using sparse subspace clustering.
problem Local attribution methods fail to identify coherent model behavior across samples.
method Sparse Subspace Clustering (SSCC) for concept discovery.
result Empirically validated method for various image classification tasks.
Sparse neural networks can match dense models on Lipschitz functions.
problem Sparse networks are more efficient but lack theoretical guarantees.
method Formal model of sparse networks, LSH-based routing function, Lipschitz function approximation.
result Sparse networks can approximate dense networks on Lipschitz functions.
The paper analyzes LIME for tabular data and proves its behavior in large samples.
problem Understanding the behavior of LIME in tabular data settings.
method Theoretical analysis of LIME's behavior in tabular data, proving its properties in the large sample limit.
result LIME provides explanations proportional to the coefficients of the function in linear cases, but can produce misleading explanations for partition-based models.
Regularization improves stability and consistency of sparse autoencoders.
problem Varying features across random seeds and training choices in SAEs.
method Added L1 or L2 penalties on encoder and decoder weights.
result L2 regularization increases cross-seed feature consistency.
Paper hypothesizes MLP layers in LLMs can be approximated by sparse Mixture of Experts.
problem Understanding dense MLP layers in LLMs.
method Theoretical connection between MoE models and SAE structure in activation space.
result MLP layers in LLMs can be well approximated by sparse Mixture of Experts.
The interaction between transitivity and sparsity, two common features in empirical networks, implies that there are local regions of large sparse networks that are dense. We call this the blessing of transitivity and it has consequences for both modeling and inference. Extant research suggests that statistical inferen…
Temporal Functional Circuits explain KAN forecasts with interpretable edge functions.
problem Lack of mechanistic explanations in KAN forecasting.
method Transform KAN edge functions into faithful, temporally grounded explanations using a gated residual KAN.
result Gated KAN achieves lower MSE than linear-only models on regime-switching signals.
Predictive modeling applications increasingly use data representing people's behavior, opinions, and interactions. Fine-grained behavior data often has different structure from traditional data, being very high-dimensional and sparse. Models built from these data are quite difficult to interpret, since they contain man…
MoreauGrad interprets neural nets robustly and sparsely.
problem Lack of robustness and sparsity in gradient-based interpretation methods.
method Moreau envelope for smooth and robust interpretation, combined with L1 regularization for sparsity.
result MoreauGrad provides a smooth, robust, and sparse explanation of neural nets.
AdaBoost improves binary classification in robust one-bit compressed sensing with adversarial errors.
problem Binary classification in robust one-bit compressed sensing with adversarial errors.
method AdaBoost and max-ℓ1-margin-classifier approach, with convergence rates improved under certain feature conditions. result Improved convergence rates and explanation for harmless interpolating adversarial noise.
SAGE-FIN detects financial fraud using GNNs and Granger causality.
problem Detecting fraud in financial networks with limited labeled data and lack of explainability.
method Semi-supervised GNN approach with Granger causal explanations.
result SAGE-FIN outperforms on real-world financial network dataset with explainable flagged items.
New method solves sparse PCA for multiple components efficiently.
problem Sparse PCA for multiple orthogonal components.
method Reformulates orthogonality as rank constraints, uses semidefinite relaxations and bounds.
result Exact solutions with near-optimal variance explained and orthogonality.
In environments with continuous state and action spaces, state-of-the-art actor-critic reinforcement learning algorithms can solve very complex problems, yet can also fail in environments that seem trivial, but the reason for such failures is still poorly understood. In this paper, we contribute a formal explanation of…
Sparse connectivity improves generalization in neural networks below the Edge of Stability.
problem Generalization guarantees for fully-connected networks fail at the Edge of Stability.
method Analyzed sparse connectivity's impact on generalization in two-layer ReLU networks.
result Sparse connectivity changes the effective constraint, leading to non-vacuous generalization bounds.
Many methods to explain black-box models, whether local or global, are additive. In this paper, we study global additive explanations for non-additive models, focusing on four explanation methods: partial dependence, Shapley explanations adapted to a global setting, distilled additive explanations, and gradient-based e…
Pruning at initialization fails to find sparse subnetworks, revealing information-theoretic barriers.
problem Difficulty in finding sparse subnetworks without training the full model.
method Analysis of effective parameter count and mutual information between sparsity mask and data.
result Pruning at initialization cannot find sparse subnetworks due to high mutual information.
RFM reduces feature space for linear models, improving sparse recovery.
problem Sparse linear regression and low-rank matrix recovery.
method Recursive Feature Machines (RFM) that alternates between reweighting feature vectors by AGOP and learning prediction function.
result RFM generalizes IRLS and outperforms deep linear networks.
GEMSS discovers multiple sparse solutions in high-dimensional data.
problem Identifying multiple sparse feature combinations in high-dimensional, underdetermined systems.
method GEMSS (Gaussian Ensemble for Multiple Sparse Solutions) uses a structured spike-and-slab prior, mixture of Gaussians, and Jaccard-based penalty to optimize a single objective function via stochastic gradient descent.
result GEMSS consistently outperforms five feature selection methods on 128 experiments and real-world datasets.
AXE evaluates explanations to avoid misleading Rashomon set model selection.
problem Evaluating explanations for Rashomon set models to avoid false selection.
method Proposed AXE method to evaluate explanation quality.
result AXE detects adversarial fairwashing with 100% success rate.
Explaining the prediction of deep neural networks (DNNs) and semantic image compression are two active research areas of deep learning with a numerous of applications in decision-critical systems, such as surveillance cameras, drones and self-driving cars, where interpretable decision is critical and storage/network ba…
The paper proposes criteria and methods for evaluating and aggregating feature-based model explanations.
problem Lack of quantitative evaluation criteria for feature-based model explanations.
method Developed quantitative evaluation criteria (low sensitivity, high faithfulness, low complexity), devised a framework for aggregation, and derived a new aggregate Shapley value explanation function.
result A new aggregate Shapley value explanation function that minimizes sensitivity.
New definition reveals encoding explanations that retain predictive power.
problem Challenges in evaluating and identifying encoding explanations.
method Developed a definition of encoding based on conditional dependence.
result Existing evaluation scores do not rank non-encoding explanations correctly, but STRIPE-X does.
New Bayesian method for sparse multidimensional item response theory.
problem Sparse interpretable explanations for questionnaire data.
method Bayesian EM algorithm for sparse factor loadings.
result Reliable recovery of factor dimensionality and latent structure.
Counterfactual explanations can be obtained by identifying the smallest change made to a feature vector to qualitatively influence a prediction; for example, from 'loan rejected' to 'awarded' or from 'high risk of cardiovascular disease' to 'low risk'. Previous approaches often emphasized that counterfactuals should be…
GRANITE unifies feature-based explanation methods to reduce disagreement.
problem Disagreement among feature-based explanation methods.
method GRANITE partitions feature space into regions minimizing interaction and distribution influences.
result Unified and consistent feature explanations.
A central goal of neuroscience is to understand how activity in the nervous system is related to features of the external world, or to features of the nervous system itself. A common approach is to model neural responses as a weighted combination of external features, or vice versa. The structure of the model weights c…
Differentially private algorithms protect model explanations from leaking training data.
problem Model explanations can leak training data, compromising privacy.
method Adaptive differentially private gradient descent algorithm to produce accurate, private explanations.
result Privacy amplification and reduction of overall privacy loss on explanation data.
We consider objective evaluation measures of saliency explanations for complex black-box machine learning models. We propose simple robust variants of two notions that have been considered in recent literature: (in)fidelity, and sensitivity. We analyze optimal explanations with respect to both these measures, and while…