XAI

Science and technology for the eXplanation of AI decision making

XAI project focuses on the urgent open challenge of how to construct meaningful explanations of opaque AI/ML systems, introducing the local-to-global framework for black box explanation, articulated along three lines: a) the language for expressing; b) the inference of local explanations; c), the bottom-up generalization of many local explanations into simple global ones

Project Information

Grant agreement ID: 834756

Call: ERC-2018-ADG

Total cost: 2 500 000€

EU Contribution: 2 500 000€

Principal investigator: Fosca Giannotti

Email: fosca.giannotti @ isti.cnr.it

An intertwined line of research will investigate i) causal explanation models that capture the causal relationships among the variables and the decision, and ii) mechanistic/physical models that capture the detailed data generation behavior behind specific deep learning models.

This project will also develop: (1) an explanation infrastructure for benchmarking, equipped with platforms for the users' assessment of the explanations; (2) an ethical-legal framework, in compliance with the provisions of the GDPR; and (3) a repertoire of case studies in explanation-by-design, mainly focused on health and fraud detection applications.

Partners​

CNR

CNR

National Research Council

Unipi

University of Pisa

Department of Computer Science

Normale

Scuola Normale Superiore

The 5 XAI
research lines

1

Local-to-global paradigm for explanation by design

2

From statistical to causal and mechanistic, physical explanations

3

XAI
platform

4

Ethical/legal framework for explanation

5

Case
studies

News

Events, tutorials, round tables, conferences and more...

Entry Image

XAI.IT

November 25-26th 2020

Italian Workshop on XAI with Dino Pedreschi as invited speaker

Entry Image

TAILOR

September 5th 2019

Foundations of Trustworthy AI - Integrating Learning, Optimization and Reasoning

Entry Image

Afirm

January 27-31th 2020

ACM SIGIR/SIGKDD Africa Summer School on Machine Learning for Data Mining and Search