POPL 2024
Sun 14 - Sat 20 January 2024 London, United Kingdom
Tue 16 Jan 2024 11:20 - 11:40 at Marconi Room - Session 6: Abstract Interpretation Chair(s): Xavier Rival

We study how a symbolic representation for support vector machines (SVMs) specified by means of abstract interpretation can be exploited for: (1) enhancing the interpretability of SVMs through a novel feature importance measure, called abstract feature importance (AFI), that does not depend in any way on a given dataset or the accuracy of the SVM and is very fast to compute; and (2) certifying individual fairness of SVMs and producing concrete counterexamples when this verification fails. We implemented our methodology and we empirically showed its effectiveness on SVMs based on linear and nonlinear (polynomial and radial basis function) kernels. Our experimental results prove that, independently of the accuracy of the SVM, our AFI measure correlates much strongly with stability of the SVM to feature perturbations than major feature importance measures available in machine learning software such as permutation feature importance, therefore providing better insight into the trustworthiness of SVMs.

Tue 16 Jan

Displayed time zone: London change

11:00 - 12:30
Session 6: Abstract InterpretationVMCAI at Marconi Room
Chair(s): Xavier Rival Inria; ENS; CNRS; PSL University
11:00
20m
Talk
Formal Runtime Error Detection During Development in the Automotive Industry
VMCAI
Jesko Hecking-Harbusch Bosch Research, Jochen Quante Bosch Research, Maximilian Schlund Bosch Research
Pre-print
11:20
20m
Talk
Abstract Interpretation-Based Feature Importance for Support Vector Machines
VMCAI
Abhinandan Pal University of Birmingham, Francesco Ranzato University of Padova, Caterina Urban Inria & École Normale Supérieure | Université PSL, Marco Zanella University of Padova, Italy
11:40
20m
Talk
Generation of Violation Witnesses by Under-Approximating Abstract Interpretation
VMCAI
Marco Milanese Sorbonne University, Antoine Miné Sorbonne Université
DOI Pre-print
12:00
20m
Talk
Correctness Witness Validation by Abstract Interpretation
VMCAI
Simmo Saan University of Tartu, Estonia, Michael Schwarz Technische Universität München, Julian Erhard Technical University of Munich, Helmut Seidl Technische Universität München, Sarah Tilscher Technische Universität München, Vesal Vojdani University of Tartu
DOI Pre-print