publication . Part of book or chapter of book . Conference object . Preprint . Article . 2020

Black box explanation by learning image exemplars in the latent feature space

Riccardo Guidotti; Anna Monreale; Stan Matwin; Dino Pedreschi;
Open Access English
  • Published: 27 Jan 2020
  • Country: Italy
Abstract
We present an approach to explain the decisions of black box models for image classification. While using the black box to label images, our explanation method exploits the latent feature space learned through an adversarial autoencoder. The proposed method first generates exemplar images in the latent feature space and learns a decision tree classifier. Then, it selects and decodes exemplars respecting local decision rules. Finally, it visualizes them in a manner that shows to the user how the exemplars can be modified to either stay within their class, or to become counter-factuals by "morphing" into another class. Since we focus on black box decision systems ...
Subjects
ACM Computing Classification System: ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION
Fields of Science and Technology classification: 02 engineering and technology0202 electrical engineeringelectronic engineering
free text keywords: Explainable AI, Adversarial autoencoder, Image exemplars, Explainable AI, Adversarial autoencoder, Image exemplars, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Computer Vision and Pattern Recognition (cs.CV), Machine Learning (cs.LG), FOS: Computer and information sciences, Contextual image classification, Pattern recognition, Decision rule, Morphing, Artificial intelligence, business.industry, business, Black box, Stability (learning theory), Feature vector, Autoencoder, Decision tree learning, Computer science
Communities
Communities with gateway
OpenAIRE Connect image
Other Communities
  • SoBigData
Funded by
EC| SoBigData
Project
SoBigData
SoBigData Research Infrastructure
  • Funder: European Commission (EC)
  • Project Code: 654024
  • Funding stream: H2020 | RIA
,
NSERC
Project
  • Funder: Natural Sciences and Engineering Research Council of Canada (NSERC)
,
EC| PRO-RES
Project
PRO-RES
PROmoting integrity in the use of RESearch results
  • Funder: European Commission (EC)
  • Project Code: 788352
  • Funding stream: H2020 | CSA
,
EC| Track and Know
Project
Track and Know
Big Data for Mobility Tracking Knowledge Extraction in Urban Areas
  • Funder: European Commission (EC)
  • Project Code: 780754
  • Funding stream: H2020 | RIA
,
EC| AI4EU
Project
AI4EU
A European AI On Demand Platform and Ecosystem
  • Funder: European Commission (EC)
  • Project Code: 825619
  • Funding stream: H2020 | RIA
Download fromView all 5 versions
Open Access
ISTI Open Portal
Conference object . 2020
Open Access
http://arxiv.org/pdf/2002.0374...
Part of book or chapter of book
Provider: UnpayWall
Closed Access
https://doi.org/10.1007/978-3-...
Part of book or chapter of book . 2020
Provider: Crossref
Any information missing or wrong?Report an Issue