Abstract Interpretation-Based Feature Importance for SVMs
This work addresses the interpretability and trustworthiness of SVMs for users in machine learning, offering a fast, dataset-independent feature importance measure and verification for stability, though it is incremental as it builds on abstract interpretation techniques.
The authors tackled the problem of interpreting and verifying support vector machines (SVMs) by proposing a symbolic representation using abstract interpretation, resulting in a novel feature importance measure (AFI) that correlates more strongly with SVM stability to feature perturbations than existing methods, as demonstrated empirically on linear and non-linear kernels.
We propose a symbolic representation for support vector machines (SVMs) by means of abstract interpretation, a well-known and successful technique for designing and implementing static program analyses. We leverage this abstraction in two ways: (1) to enhance the interpretability of SVMs by deriving a novel feature importance measure, called abstract feature importance (AFI), that does not depend in any way on a given dataset of the accuracy of the SVM and is very fast to compute, and (2) for verifying stability, notably individual fairness, of SVMs and producing concrete counterexamples when the verification fails. We implemented our approach and we empirically demonstrated its effectiveness on SVMs based on linear and non-linear (polynomial and radial basis function) kernels. Our experimental results show that, independently of the accuracy of the SVM, our AFI measure correlates much more strongly with the stability of the SVM to feature perturbations than feature importance measures widely available in machine learning software such as permutation feature importance. It thus gives better insight into the trustworthiness of SVMs.