Skip to content

[AAAI 2020] Latent Opinions Transfer Network for Target-Oriented Opinion Words Extraction

Notifications You must be signed in to change notification settings

1429904852/LOTN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Aug 23, 2022
43f6793 · Aug 23, 2022

History

10 Commits
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 22, 2022
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 23, 2022
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020
Aug 9, 2020

Repository files navigation

LOTN

Codes and datasets for our AAAI 2022 paper: "Latent Opinions Transfer Network for Target-Oriented Opinion Words Extraction"

Overview

  • In tackling the problem of insufficient annotated data, we are the first to propose transferring latent opinion knowledge from resource-rich review sentiment classification datasets to the low-resource task of TOWE.
  • To transfer opinion information effectively, we propose a Latent Opinions Transfer Network (LOTN) model that obtains latent opinion words from a sentiment classification model and integrates them into TOWE via an auxiliary learning signal.

Dependencies

  • python=3.5
  • numpy=1.14.2
  • tensorflow==1.9

Usage

Step1: pretrained

sh run_pre_trained.sh

step2: transfer

  • Softmax
sh run_LOTN.sh
python eval_LOTN.py
  • CRF
sh run_LOTN.sh
python eval_LOTN_crf.py

Citation

@inproceedings{DBLP:conf/aaai/WuZDHC20,
  author    = {Zhen Wu and Fei Zhao and Xin{-}Yu Dai and Shujian Huang and Jiajun Chen},
  title     = {Latent Opinions Transfer Network for Target-Oriented Opinion Words Extraction},
  booktitle = {The Thirty-Fourth {AAAI} Conference on Artificial Intelligence, {AAAI} 2020, New York, NY, USA, February 7-12, 2020},
  pages     = {9298--9305},
  year      = {2020}
}

if you have any questions, please contact me [email protected].

About

[AAAI 2020] Latent Opinions Transfer Network for Target-Oriented Opinion Words Extraction

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published