We propose an interactive machine learning framework where the machine questions the user feedback when it realizes it is inconsistent with the knowledge previously accumulated. The key idea is that the machine uses its available knowledge to check the correctness of its own and the user labeling. The proposed architecture and algorithms run through a series of modes with progressively higher confdence and features a confict resolution component. The proposed solution is tested in a project on university student life where the goal is to recognize tasks like user location and transportation mode from sensor data. The results highlight the unexpected extreme pervasiveness of annotation mistakes and the advantages provided by skeptical learning.

Dealing with Mislabeling via Interactive Machine Learning / Zhang, Wanyi; Passerini, Andrea; Giunchiglia, Fausto. - In: KI - KÜNSTLICHE INTELLIGENZ. - ISSN 0933-1875. - 34:2(2020), pp. 271-278. [10.1007/s13218-020-00630-5]

Dealing with Mislabeling via Interactive Machine Learning

Wanyi Zhang;Andrea Passerini;Fausto Giunchiglia
2020-01-01

Abstract

We propose an interactive machine learning framework where the machine questions the user feedback when it realizes it is inconsistent with the knowledge previously accumulated. The key idea is that the machine uses its available knowledge to check the correctness of its own and the user labeling. The proposed architecture and algorithms run through a series of modes with progressively higher confdence and features a confict resolution component. The proposed solution is tested in a project on university student life where the goal is to recognize tasks like user location and transportation mode from sensor data. The results highlight the unexpected extreme pervasiveness of annotation mistakes and the advantages provided by skeptical learning.
2020
2
Zhang, Wanyi; Passerini, Andrea; Giunchiglia, Fausto
Dealing with Mislabeling via Interactive Machine Learning / Zhang, Wanyi; Passerini, Andrea; Giunchiglia, Fausto. - In: KI - KÜNSTLICHE INTELLIGENZ. - ISSN 0933-1875. - 34:2(2020), pp. 271-278. [10.1007/s13218-020-00630-5]
File in questo prodotto:
File Dimensione Formato  
10.1007_s13218-020-00630-5 (1).pdf

Solo gestori archivio

Descrizione: Author's personal copy
Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 985.61 kB
Formato Adobe PDF
985.61 kB Adobe PDF   Visualizza/Apri
Zhang2020_Article_DealingWithMislabelingViaInter.pdf

Solo gestori archivio

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 942.85 kB
Formato Adobe PDF
942.85 kB Adobe PDF   Visualizza/Apri
2019_KI___Skeptical_learning.pdf

accesso aperto

Tipologia: Post-print referato (Refereed author’s manuscript)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 460.26 kB
Formato Adobe PDF
460.26 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11572/268929
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? 1
social impact