jchanxtarov / arXivNotes

repository to research & memo about the machine learning articles

MIT License

1 stars 1 forks source link

VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback #17

Open jchanxtarov opened 3 years ago

jchanxtarov commented 3 years ago

Basic Information

Authors: Ruining He, Julian McAuley
Date: Oct 2015
Published By: AAAI'16

Link

https://arxiv.org/abs/1510.01784

Overview

In the proposed model, visual features for the task of personalized ranking on implicit feedback datasets are incorporated.
(Key point) With the existing MF, it is not possible to learn the expression of (cold) items that are bought only in small numbers. -> The embedded information is also learned for the visual image.
Structure:

Others

Hybrid models (context-aware recommendations) to solve cold start problem
- from text: (Bao, Fang, and Zhang, 2014)
- user's physical location: (Qiao et al., 2014)
- season or temperature: (Brown, Bovey, and Chen, 1997)
- (This study) visual appearance

Reference (for understanding)

jchanxtarov commented 3 years ago

実装は簡単そう，今晩書いてみる

jchanxtarov commented 3 years ago

できた https://github.com/jchanxtarov/vbpr