Knowledge-based Transfer Learning Explanation

Jiaoyan Chen Freddy Lecue 1, 2 Jeff Pan Ian Horrocks 3 Huajun Chen
2 WIMMICS - Web-Instrumented Man-Machine Interactions, Communities and Semantics
CRISAM - Inria Sophia Antipolis - Méditerranée , Laboratoire I3S - SPARKS - Scalable and Pervasive softwARe and Knowledge Systems
Abstract : Machine learning explanation can significantly boost machine learning's application in decision making, but the usability of current methods is limited in human-centric explanation, especially for transfer learning, an important machine learning branch that aims at utilizing knowledge from one learning domain (i.e., a pair of dataset and prediction task) to enhance prediction model training in another learning domain. In this paper , we propose an ontology-based approach for human-centric explanation of transfer learning. Three kinds of knowledge-based explanatory evidence, with different granularities, including general factors, particular narrators and core contexts are first proposed and then inferred with both local ontologies and external knowledge bases. The evaluation with US flight data and DB-pedia has presented their confidence and availability in explaining the transferability of feature representation in flight departure delay forecasting.
Type de document :
Communication dans un congrès
Principles of Knowledge Representation and Reasoning: Proceedings of the Sixteenth International Conference, Oct 2018, Tempe, United States
Liste complète des métadonnées

https://hal.inria.fr/hal-01934907
Contributeur : Freddy Lecue <>
Soumis le : lundi 26 novembre 2018 - 12:54:48
Dernière modification le : jeudi 29 novembre 2018 - 01:19:56

Fichier

KR2018.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01934907, version 1

Collections

Citation

Jiaoyan Chen, Freddy Lecue, Jeff Pan, Ian Horrocks, Huajun Chen. Knowledge-based Transfer Learning Explanation. Principles of Knowledge Representation and Reasoning: Proceedings of the Sixteenth International Conference, Oct 2018, Tempe, United States. 〈hal-01934907〉

Partager

Métriques

Consultations de la notice

10

Téléchargements de fichiers

6