ZeroFPR_SVDD

Detection of safety regions (zero False Positive) via Support Vector Data Description (SVDD).

ZeroFPR_SVDD

Safety regions research is a well-known task for ML and the main focus is to avoid false positives, i.e., including in the safe region unsafe points. In this repository, two methods for the research of zero FPR regions are proposed: the first one is based simply on the reduction of the SVDD radius until only safe points are enclosed in the SVDD shape; the second one instead performs successive iterations of the SVDD on the safe region until there are no more negative points.

Zero_FPR_exe.m is the example code of the the algorithm, while ZeroFPR_SVDD.m is the code of the main algorithm.

Cite As -) Plain text A. Carlevaro and M. Mongelli, "A New SVDD Approach to Reliable and Explainable AI," in IEEE Intelligent Systems, vol. 37, no. 2, pp. 55-68, 1 March-April 2022, doi: 10.1109/MIS.2021.3123669.

-) BibTex @ARTICLE{9594676, author={Carlevaro, Alberto and Mongelli, Maurizio}, journal={IEEE Intelligent Systems}, title={A New SVDD Approach to Reliable and Explainable AI}, year={2022}, volume={37}, number={2}, pages={55-68}, doi={10.1109/MIS.2021.3123669}}

Author ORCID: https://orcid.org/0000-0002-7206-5511