logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-04122019-110146


Tipo di tesi
Tesi di laurea magistrale
Autore
AFFOLTER, RICCARDO
Indirizzo email
riccardo.affolter@gmail.com
URN
etd-04122019-110146
Titolo
Xlib - Explanation Library for Black Box Models
Dipartimento
INFORMATICA
Corso di studi
INFORMATICA PER L'ECONOMIA E PER L'AZIENDA (BUSINESS INFORMATICS)
Relatori
relatore Dott. Guidotti, Riccardo
relatore Dott.ssa Monreale, Anna
relatore Giannotti, Fosca
controrelatore Milazzo, Paolo
Parole chiave
  • bias
  • data mining
  • images
  • machine learning
  • python library
Data inizio appello
03/05/2019
Consultabilità
Non consultabile
Data di rilascio
03/05/2089
Riassunto
Many machine learning algorithms are becoming a useful computational tool to find answers to support decisions, but they are strongly criticized for the lack of interpretability in relation to the predictions they provide. In favour of the compromise between accuracy and interpretability, many proposals have been found with many different approaches already developed.
This thesis focuses on the creation of a stable and usable platform that works with some explanation methods existing in literature. According to this goal it is necessary to discover proposed methods dealing with this topic and insert them into a platform. This explainable system gives transparency in the decision policies of some predictive models that might be used by experts and non-experts in the field of machine learning. To support developers, researchers and practitioners, Xlib have to allow easy and standardized access to a wide variety of algorithms integrated into a common framework, to evaluate, compare and visualize the results they provide.
File