alpha-TCAV: A Unified Framework for Testing with Concept Activation Vectors
stat.ML, cs.AI, cs.LG, math.PR
Submitted: 2026-05-15
Updated: 2026-10-12
Terminology
Sources
- Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
- Concept-Based Explainable Artificial Intelligence: Metrics and Benchmarks
- Post-hoc Concept Bottleneck Models
- Visual-TCAV: Concept-based Attribution and Saliency Maps for Post-hoc Explainability in Image Classification
- Concept Bottleneck Large Language Models
- Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering
- On The Variability of Concept Activation Vectors
- Two friendly proofs of the Berry-Esseen theorem
- Looking Deeper into Tabular LIME
- A Unified Framework for Provably Efficient Algorithms to Estimate Shapley Values
- Semi-analytical approximations to statistical moments of sigmoid and softmax mappings of normal variables
- Large Dimensional Analysis and Improvement of Multi Task Learning
- Concentration of Measure and Large Random Matrices with an application to Sample Covariance Matrices
- Characterization of Gaussian Universality Breakdown in High-Dimensional Empirical Risk Minimization
Related papers
- Behavior of prediction performance metrics with rare events
- Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
- A Posterior-Dynamics Framework for Imaging Inverse Problems with Pretrained Diffusion Priors
- Efficient quadratic entropy with distance sketches
- Embedding-Bias in Conditional Independence Testing
- From Log-Odds to Shapley Values: An Explanatory Geometry for the Weighted Naive Bayes Classifier