Skip to content
/ AMR Public

This is our official implementation for the paper: Jinhui Tang, Xiaoyu Du, Xiangnan He, Fajie Yuan, Qi Tian, and Tat-Seng Chua, Adversarial Training Towards Robust Multimedia Recommender System.

Notifications You must be signed in to change notification settings

duxy-me/AMR

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Adversarial Training Towards Robust Multimedia Recommender System

Appending adversarial training on multimedia features enhances the performance of multimedia recommender system.

This is our official implementation for the paper:

Jinhui Tang, Xiangnan He, Xiaoyu Du, Fajie Yuan, Qi Tian, and Tat-Seng Chua, Adversarial Training Towards Robust Multimedia Recommender System.

If you use the codes, please cite our paper. Thanks!

Requirements

  • Tensorflow 1.7
  • numpy, scipy

Quick Start

figure.png

  1. Data

    • f_resnet.npy Deep image features extracted with Resnet. The $i$-th row indicates the $i$-th item feature.
    • pos.txt The training samples used in training process. The numbers $u$ and $i$ in each row indicate an interaction between user $u$ and item $i$.
    • neg.txt The test samples used in testing process. The first number of row $u$ is the only positive sample in test, the following numbers of row $u$ are the negative samples for user $u$.
  2. Pretrained VBPR The pretrained VBPR is stored in weights/best-vbpr.npy

  3. Traing AMR

    bash run.sh
    

    The training logs are stored in logs

Source Files

Source files are stored in src/.

  • main.py. The main entrance of the program.

  • solver/*. The solvers managing the training process.

  • model/*. The models.

  • dataset/*. The data readers.

About

This is our official implementation for the paper: Jinhui Tang, Xiaoyu Du, Xiangnan He, Fajie Yuan, Qi Tian, and Tat-Seng Chua, Adversarial Training Towards Robust Multimedia Recommender System.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published