Skip to content

Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding

Notifications You must be signed in to change notification settings

MiuLab/Time-SLU

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Time-SLU: Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding

An implementation of the ASRU 2017 paper: Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding.

Content

Data

the data used in the paper is DSTC4

Requirements

Tensorflow ver1.2 CUDNN ver5.1 Python 2.7

Usage

  • Change the path in slu_preprocess.py line 29 to your custom GloVe file path.
  • bash run.sh will reproduce log files for every entry in table 1.
  • python2.7 calculate.py will calculate the average of log files for each entry in Table 1.

Reference

Main papers to be cited

@inproceedings{chen2017dynamic,
  author    = {Po-Chun Chen and Ta-Chung Chi and Shang-Yu Su and Yun-Nung Chen},
  title	    = {Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding},
  booktitle = {Proceedings of 2017 IEEE Workshop on Automatic Speech Recognition and Understanding},
  year	    = {2017},
  address   = {Okinawa, Japan}
}

About

Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published