The study evaluates how well local explanations align with model predictions.
problem Capturing the faithfulness of local explanations to model predictions.
method Introducing consistency and sufficiency as properties, and developing quantitative measures and estimators.
result Quantitative measures of consistency and sufficiency depend on test-time data distribution.
Measures faithfulness of LLM explanations to reveal hidden biases and misleading claims.
problem LLM explanations can misrepresent the model's reasoning process, leading to over-trust and misuse.
method Defines faithfulness in terms of concept influence and uses counterfactuals and Bayesian models to estimate it.
result Can quantify and discover interpretable patterns of unfaithfulness in LLM explanations.
ID-ExpO fine-tunes neural networks for more faithful explanations.
problem Improving the faithfulness of explanations for complex machine learning models.
method Differentiable insertion/deletion metric-aware regularizers for optimization.
result Fine-tuned predictors produce more faithful explanations.
The paper proposes new interpretability paradigms to improve model faithfulness.
problem Improving the accuracy of explanations for complex models.
method Examining and evolving existing paradigms, proposing new models.
result Three new paradigms for interpretability are presented.
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.
LIMEtree offers faithful explanations for multiple classes in predictive models.
problem Generating explanations for several classes can be difficult due to conflicting evidence.
method LIMEtree uses multi-output regression trees for consistent and faithful explanations of multiple classes.
result LIMEtree provides diverse explanation types and outperforms LIME in evaluations.
EXAGREE selects a stakeholder-aligned model to reduce conflicting explanations in machine learning.
problem Conflicting explanations from different attribution methods limit the adoption of machine learning models in safety-critical domains.
method EXAGREE is a two-stage framework that selects a Stakeholder-Aligned Explanation Model (SAEM) from a set of similar-performing models, maximizing Stakeholder-Machine Agreement (SMA).
result EXAGREE achieves simultaneous gains in faithfulness, plausibility, and fairness over baselines while preserving task accuracy.
Model interpretability is an increasingly important component of practical machine learning. Some of the most common forms of interpretability systems are example-based, local, and global explanations. One of the main challenges in interpretability is designing explanation systems that can capture aspects of each of th…
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.
PiNets provide faithful explanations for neural networks.
problem Lack of true explanations for neural network predictions.
method Pointwise-interpretable Networks (PiNets) that form linear models instance-wise.
result PiNets offer explanations that are meaningful, aligned, robust, and sufficient.
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.
We introduce a new model-agnostic explanation technique which explains the prediction of any classifier called CLE. CLE gives an faithful and interpretable explanation to the prediction, by approximating the model locally using an interpretable model. We demonstrate the flexibility of CLE by explaining different models…
Study evaluates feature ranking methods' faithfulness in ML models, improving with dimensionality reduction.
problem Quantifying and improving the faithfulness of feature ranking methods in ML models.
method Evaluation of multiple feature ranking methods, including SHAP, LIME, ALE variance, and LR coefficients, using permutation importance as a baseline.
result Dimensionality reduction improves the faithfulness of feature ranking methods, making permutation importance the most faithful method.
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.
XCM improves MTS classification with explainable deep learning.
problem Lack of explainable deep learning models for MTS classification.
method XCM is a compact CNN that extracts variable and timestamp information directly from input data.
result XCM outperforms state-of-the-art MTS classifiers on large and small datasets.
Most recent work on interpretability of complex machine learning models has focused on estimating a posteriori explanations for previously trained models around specific predictions. Self-explaining models where interpretability plays a key role already during learning have received much less atte…
Explainable Artificial Intelligence (XAI)has received a great deal of attention recently. Explainability is being presented as a remedy for the distrust of complex and opaque models. Model agnostic methods such as LIME, SHAP, or Break Down promise instance-level interpretability for any complex machine learning model. …
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.
Attack reveals model details from counterfactual explanations.
problem Extracting model details from counterfactual explanations.
method Adversary uses counterfactual explanations to build high-fidelity model.
result High-fidelity and high-accuracy model extraction possible.
ALMANACS benchmarks explainability methods on simulatability.
problem Evaluating the effectiveness of explainability methods for language models.
method ALMANACS is a simulatability benchmark that evaluates explainability methods on twelve safety-relevant topics.
result No explainability method outperforms the explanation-free control across all topics.
Proposes a method for explaining black-box models with nested feature attributions.
problem Making black-box models transparent and trustworthy.
method Model-agnostic local explanation method exploiting nested feature structure and consistency property.
result Accurate and consistent HiFAs and LoFAs estimated using fewer model queries.
OrphicX generates causal explanations for GNNs by isolating latent causal factors.
problem Generating interpretable causal explanations for complex graph neural networks.
method Develops a generative model and objective function to isolate latent causal factors, maximizing information flow.
result OrphicX effectively identifies causal semantics, significantly outperforming alternatives.
PACE explains ViTs by modeling patch-level concept distributions, surpassing existing methods.
problem Lack of trustworthy post-hoc explanations for Vision Transformers (ViTs)
method Variational Bayesian explanation framework (PACE)
result PACE surpasses state-of-the-art methods in meeting desiderata for ViT explanations.
The paper explores methods to explain decisions of deep learning models by faithfully reproducing their training data views.
problem Explaining decisions of complex deep learning models trained on large datasets.
method Data view extraction through hill-climbing and GAN-driven approaches, followed by creation of shadow models for explanation.
result Shadow models based on faithfully reproduced data views are effective for explaining decisions of blackbox deep learning models.
Modified BP attribution methods often ignore later layers' information, leading to misleading explanations.
problem Misleading explanations from modified BP methods ignoring later layers' information.
method Analysis of 9 modified BP methods including Deep Taylor Decomposition, LRP, Excitation BP, PatternAttribution, DeepLIFT, Deconv, RectGrad, Guided BP.
result Only DeepLIFT does not ignore later layers' information, providing a faithful explanation.
Active learning has long been a topic of study in machine learning. However, as increasingly complex and opaque models have become standard practice, the process of active learning, too, has become more opaque. There has been little investigation into interpreting what specific trends and patterns an active learning st…
Despite outstanding contribution to the significant progress of Artificial Intelligence (AI), deep learning models remain mostly black boxes, which are extremely weak in explainability of the reasoning process and prediction results. Explainability is not only a gateway between AI and society but also a powerful tool t…
Neural networks are among the most accurate supervised learning methods in use today, but their opacity makes them difficult to trust in critical applications, especially when conditions in training differ from those in test. Recent work on explanations for black-box models has produced tools (e.g. LIME) to show the im…
Improves recommender system explainability by clarifying representation learning.
problem Lack of explainability in recommender systems.
method Proposes a novel explainable recommendation model by improving transparency in representation learning.
result The proposed model learns interpretable representations that are faithful to explanations.
Tree-based machine learning models such as random forests, decision trees, and gradient boosted trees are the most popular non-linear predictive models used in practice today, yet comparatively little attention has been paid to explaining their predictions. Here we significantly improve the interpretability of tree-bas…
We study complex hyperbolic disc bundles over closed orientable surfaces that arise from discrete and faithful representations H_n->PU(2,1), where H_n is the fundamental group of the orbifold S^2(2,...,2) and thus contains a surface group as a subgroup of index 2 or 4. The results obtained provide the first complex hyp…
New method R-LOCO improves local feature importance analysis.
problem Local attribution methods fail to accurately identify important features.
method R-LOCO segments input space into regions and applies global methods within.
result R-LOCO delivers more accurate local attributions.
While neural networks have acted as a strong unifying force in the design of modern AI systems, the neural network architectures themselves remain highly heterogeneous due to the variety of tasks to be solved. In this chapter, we explore how to adapt the Layer-wise Relevance Propagation (LRP) technique used for explain…
Paper identifies problematic baselines in Shapley value explanations and proposes a reweighting mechanism.
problem Identifying and addressing the suboptimality of baselines in Shapley value feature importance analysis.
method Analyzed suboptimality of baselines, identified problematic baseline, generalized uninformativeness, and designed a reweighting mechanism.
result Proposed uncertainty-based reweighting mechanism effectively accelerates computation and improves explanation quality.
Paper proposes counterfactual explanations for ML on multivariate time series data.
problem Lack of user trust and difficulty in debugging ML frameworks using multivariate time series data.
method Proposes a novel explainability technique for providing counterfactual explanations.
result Outperforms state-of-the-art explainability methods in metrics like faithfulness and robustness.
Despite widespread adoption, machine learning models remain mostly black boxes. Understanding the reasons behind predictions is, however, quite important in assessing trust, which is fundamental if one plans to take action based on a prediction, or when choosing whether to deploy a new model. Such understanding also pr…
CB-SLICE identifies concept-based error slices in deep learning models.
problem Systematic errors in deep learning models on specific groups.
method Concept Bottleneck Models (CBMs) and concept representations.
result CB-SLICE outperforms state-of-the-art methods in error slice identification.
FRED explains text model predictions by identifying key words and providing counterfactual examples.
problem Lack of interpretable methods for text models that are complex, lack foundations, and have unguaranteed performance.
method FRED identifies minimal influential word sets, assigns token importance, and generates counterfactual examples.
result FRED provides reliable and effective explanations for text model predictions.
TSL learns separable models to avoid signal cancellation and off-support extrapolation.
problem Signal cancellation and off-support extrapolation in additive models.
method Tensor Separation Learning (TSL) via stagewise greedy procedure with orthogonal refitting.
result TSL avoids information loss caused by marginalizing higher-order interactions.
PWSHAP provides targeted explanations for complex models.
problem Inability of black-box models to explain targeted effects in sensitive domains.
method Augments model with DAG, uses Shapley values for causal pathway identification.
result Establishes error bounds and demonstrates resolution, interpretability, and locality.
A new framework explains GNN predictions by simulating graph structure and feature changes.
problem Lack of transparency in GNN predictions hinders understanding.
method TraP2 framework using a three-layer architecture: Translation, Perturbation, and Paraphrase layers.
result TraP2 achieves 10.2% higher explanation accuracy than state-of-the-art methods.
COCKATIEL explains neural net models on NLP tasks by identifying meaningful concepts.
problem Transformer models are complex and hard to interpret.
method COCKATIEL uses NMF and sensitivity analysis to identify and rank concepts used by the model.
result COCKATIEL provides accurate and meaningful explanations without affecting model performance.
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…
ShapShift explains shifts in model predictions due to data distribution changes.
problem Prediction shifts caused by changes in input distribution.
method Subgroup Conditional Shapley Values applied to decision trees and ensembles.
result Simple, faithful, and near-complete explanations of prediction shifts across model classes.
This work introduces a new metric to assess the fidelity of surrogate models to the underlying data-generating signal.
problem The limitations of fidelity-based explanations in explainable AI.
method Introduces the linearity score λ(f) to quantify the extent of a regression network's linear decodability. result High-fidelity surrogates can underperform compared to simpler models and even linear baselines trained directly on the data.
Study on faithfulness of Burau representation for Artin-Tits groups.
problem Determining faithfulness of Burau representation for Artin-Tits groups.
method Algorithmic approach based on categorical methods.
result Burau representation is not faithful in specific cases.
Supervisory signals can help topic models discover low-dimensional data representations that are more interpretable for clinical tasks. We propose a framework for training supervised latent Dirichlet allocation that balances two goals: faithful generative explanations of high-dimensional data and accurate prediction of…
Proves Burau representation is faithful for 4-strand braids.
problem Proving the Burau representation is faithful for 4-strand braids.
method Used ideas from Moody, Long, Long-Paton, and Bigelow.
result Burau representation is faithful for n = 4.