AITM

TensorFlow implementation of Adaptive Information Transfer Multi-task (AITM) framework.
Code for the paper accepted by KDD21: Modeling the Sequential Dependence among Audience Multi-step Conversions with Multi-task Learning in Targeted Display Advertising. [https://arxiv.org/abs/2105.08489]

Note(New)

Recently we reformatted the model code as AITM_standard.py. If you want to run the model on multiple tasks (more than two), you can directly pass in the parameter --num_tasks in AITM_standard.py. But you need to configure the config.csv file to specify the size of the feature dictionary.

Requirement

python==3.6
tensorflow-gpu==1.10.1

Dataset

We use the public Ali-CCP (Alibaba Click and Conversion Prediction) dataset. [https://tianchi.aliyun.com/datalab/dataSet.html?dataId=408].

Please download and unzip the dataset first.

Split the data to train/validation/test files to run the codes directly:

python process_public_dataset.py

Example to run the model

python AITM.py --embedding_dim 5 --lr 1e-3 --early_stop 1 --lamda 1e-6 --prefix AITM --weight 0.6

The instruction of commands has been clearly stated in the codes (see the parse_args function).

Reference

If you are interested in the code, please cite our paper:

Xi D, Chen Z, Yan P, et al. Modeling the sequential dependence among audience multi-step conversions with multi-task learning in targeted display advertising[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. 2021: 3745-3755.

or in bibtex style:

@inproceedings{xi2021modeling,
  title={Modeling the sequential dependence among audience multi-step conversions with multi-task learning in targeted display advertising},
  author={Xi, Dongbo and Chen, Zhen and Yan, Peng and Zhang, Yinger and Zhu, Yongchun and Zhuang, Fuzhen and Chen, Yu},
  booktitle={Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery \& Data Mining},
  pages={3745--3755},
  year={2021}
}

Other unofficial implementations for reference:

A PyTorch implementation of multi-task recommendation models

[https://github.com/easezyc/Multitask-Recommendation-Library]

Last Update Date: Oct. 19, 2023 (UTC+8)