This is our proposed model for subtask 1 and 2 of Dialog System Technology Challenges 8 (DSTC 8). Our overall method get 3rd and 4th position of subtask1 and subtask2 according to the evaluation of contest.

Python 3.6 Tensorflow 1.4.0

Open DSTC_8_solution/Pre_trainprocess/sh_file Run bash REBED_advisor_adapt_external.sh and bash REBED_ubuntu_adapt_external.sh for advisor dataset and ubuntu dataset, respectively.

Open DSTC_8_solution/DSTC_finetune/sh_file Train scripts run_advisor_HAE.sh, run_ubuntu_HAE.sh, run_task2_REBED_adapt.sh Eval scrips eval_task1_advisor.sh, eval_task1_ubuntu.sh, eval_task2.sh

If you use the code, please cite the following paper:

@misc{gu2020pretrained, title={Pre-Trained and Attention-Based Neural Networks for Building Noetic Task-Oriented Dialogue Systems}, author={Jia-Chen Gu and Tianda Li and Quan Liu and Xiaodan Zhu and Zhen-Hua Ling and Yu-Ping Ruan}, year={2020}, eprint={2004.01940}, archivePrefix={arXiv}, primaryClass={cs.CL} }