Model-based approaches to recommendation have proven to be very accurate. Unfortunately, exploiting a latent space we miss references to the actual semantics of recommended items. In this extended abstract, we show how to initialize latent factors in Factorization Machines by using semantic features coming from a knowledge graph in order to train an interpretable model. Finally, we introduce and evaluate the semantic accuracy and robustness for the knowledge-aware interpretability of the model.

Semantic interpretability of latent factors for recommendation / Anelli, Vito Walter; Di Noia, Tommaso; Di Sciascio, Eugenio; Ragone, Azzurra; Pomo, Claudio. - ELETTRONICO. - 2441:(2019), pp. 43-44. (Intervento presentato al convegno 10th Italian Information Retrieval Workshop, IIR 2019 tenutosi a Padova, Italy nel September 16-18, 2019).

Semantic interpretability of latent factors for recommendation

Vito Walter Anelli;Tommaso Di Noia;Eugenio Di Sciascio;Claudio Pomo
2019-01-01

Abstract

Model-based approaches to recommendation have proven to be very accurate. Unfortunately, exploiting a latent space we miss references to the actual semantics of recommended items. In this extended abstract, we show how to initialize latent factors in Factorization Machines by using semantic features coming from a knowledge graph in order to train an interpretable model. Finally, we introduce and evaluate the semantic accuracy and robustness for the knowledge-aware interpretability of the model.
2019
10th Italian Information Retrieval Workshop, IIR 2019
Semantic interpretability of latent factors for recommendation / Anelli, Vito Walter; Di Noia, Tommaso; Di Sciascio, Eugenio; Ragone, Azzurra; Pomo, Claudio. - ELETTRONICO. - 2441:(2019), pp. 43-44. (Intervento presentato al convegno 10th Italian Information Retrieval Workshop, IIR 2019 tenutosi a Padova, Italy nel September 16-18, 2019).
File in questo prodotto:
File Dimensione Formato  
paper2.pdf

accesso aperto

Tipologia: Versione editoriale
Licenza: Creative commons
Dimensione 1.26 MB
Formato Adobe PDF
1.26 MB 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/11589/228466
Citazioni
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact