This code provides only DSRIM network architecture and training process. You must prepare the vectors data in order to test the network.
There is some comments in code to help understanding functions. If you have any question concerning the source code, feel free to contact the author or open a github issue.
Our network is made with Torch 7 framework. These packages is required in the codes:
- torch
- nninit
- nn
- optim
If you use this code in your work, you are welcome to cite our ICTIR 2017 paper.
@inproceedings{Nguyen:2017:DDN:3121050.3121063,
author = {Nguyen, Gia-Hung and Soulier, Laure and Tamine, Lynda and Bricon-Souf, Nathalie},
title = {DSRIM: A Deep Neural Information Retrieval Model Enhanced by a Knowledge Resource Driven Representation of Documents},
booktitle = {Proceedings of the ACM SIGIR International Conference on Theory of Information Retrieval},
series = {ICTIR '17},
year = {2017},
pages = {19--26},
publisher = {ACM},
}