GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → hsqmlzno1 → Transferable-E2E-ABSA

hsqmlzno1 / Transferable-E2E-ABSA

Licence: MIT license
Transferable End-to-End Aspect-based Sentiment Analysis with Selective Adversarial Learning (EMNLP'19)

Programming Languages

python
139335 projects - #7 most used programming language
shell
77523 projects

Labels

sequence-labelingabsaadversarial-learningdomain-adaptationaspect-based-sentiment-analysiscross-domain-sentiment

Projects that are alternatives of or similar to Transferable-E2E-ABSA

MGAN
Exploiting Coarse-to-Fine Task Transfer for Aspect-level Sentiment Classification (AAAI'19)
Stars: ✭ 44 (-29.03%)
Mutual labels:  domain-adaptation, aspect-based-sentiment-analysis, cross-domain-sentiment
adapt
Awesome Domain Adaptation Python Toolbox
Stars: ✭ 46 (-25.81%)
Mutual labels:  adversarial-learning, domain-adaptation
Tf Dann
Domain-Adversarial Neural Network in Tensorflow
Stars: ✭ 556 (+796.77%)
Mutual labels:  adversarial-learning, domain-adaptation
DAS
Code and datasets for EMNLP2018 paper ‘‘Adaptive Semi-supervised Learning for Cross-domain Sentiment Classification’’.
Stars: ✭ 48 (-22.58%)
Mutual labels:  domain-adaptation, cross-domain-sentiment
CADA
Attending to Discriminative Certainty for Domain Adaptation
Stars: ✭ 17 (-72.58%)
Mutual labels:  adversarial-learning, domain-adaptation
MemNet ABSA
No description or website provided.
Stars: ✭ 20 (-67.74%)
Mutual labels:  absa, aspect-based-sentiment-analysis
Adaptsegnet
Learning to Adapt Structured Output Space for Semantic Segmentation, CVPR 2018 (spotlight)
Stars: ✭ 654 (+954.84%)
Mutual labels:  adversarial-learning, domain-adaptation
Gvb
Code of Gradually Vanishing Bridge for Adversarial Domain Adaptation (CVPR2020)
Stars: ✭ 52 (-16.13%)
Mutual labels:  adversarial-learning, domain-adaptation
Clan
( CVPR2019 Oral ) Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptation
Stars: ✭ 248 (+300%)
Mutual labels:  adversarial-learning, domain-adaptation
Awesome Domain Adaptation
A collection of AWESOME things about domian adaptation
Stars: ✭ 3,357 (+5314.52%)
Mutual labels:  adversarial-learning, domain-adaptation
AspectBasedSentimentAnalysis
Aspect Based Sentiment Analysis is a special type of sentiment analysis. In an explicit aspect, opinion is expressed on a target(opinion target), this aspect-polarity extraction is known as ABSA.
Stars: ✭ 61 (-1.61%)
Mutual labels:  absa, aspect-based-sentiment-analysis
CrossNER
CrossNER: Evaluating Cross-Domain Named Entity Recognition (AAAI-2021)
Stars: ✭ 87 (+40.32%)
Mutual labels:  sequence-labeling, domain-adaptation
visda2019-multisource
Source code of our submission (Rank 1) for Multi-Source Domain Adaptation task in VisDA-2019
Stars: ✭ 49 (-20.97%)
Mutual labels:  domain-adaptation
IAST-ECCV2020
IAST: Instance Adaptive Self-training for Unsupervised Domain Adaptation (ECCV 2020) https://teacher.bupt.edu.cn/zhuchuang/en/index.htm
Stars: ✭ 84 (+35.48%)
Mutual labels:  domain-adaptation
Domain-Consensus-Clustering
[CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation
Stars: ✭ 85 (+37.1%)
Mutual labels:  domain-adaptation
Deep-Unsupervised-Domain-Adaptation
Pytorch implementation of four neural network based domain adaptation techniques: DeepCORAL, DDC, CDAN and CDAN+E. Evaluated on benchmark dataset Office31.
Stars: ✭ 50 (-19.35%)
Mutual labels:  domain-adaptation
BA3US
code for our ECCV 2020 paper "A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation"
Stars: ✭ 31 (-50%)
Mutual labels:  domain-adaptation
collateral-learning
Collateral Learning - Functional Encryption and Adversarial Training on partially encrypted networks
Stars: ✭ 67 (+8.06%)
Mutual labels:  adversarial-learning
SA-DL
Sentiment Analysis with Deep Learning models. Implemented with Tensorflow and Keras.
Stars: ✭ 35 (-43.55%)
Mutual labels:  aspect-based-sentiment-analysis
StyleGAN-nada
stylegan-nada.github.io/
Stars: ✭ 1,018 (+1541.94%)
Mutual labels:  domain-adaptation
View All Similar Projects ➔

Transferable-E2E-ABSA

Data and source code for our EMNLP'19 Long paper, oral, "Transferable End-to-End Aspect-based Sentiment Analysis with Selective Adversarial Learning".

Update:

Oct 6th, 2019: The experimental code (not fully clean version) has been released.

Oct 31th, 2019: The paper has been released in the arkiv.

Introduction

1) E2E-ABSA: This task aims to jointly learn aspects as well as their sentiments from user reviews, whch can be effectively formulated as an end-to-end sequence labeling problem based on the unified tagging scheme.

The unified tagging is similar to the NER tagging.

unified tag = aspect boundary tag + sentiment tag

NER tag = entity boundary tag + entity type tag

As we all know, labeling sequence data behaves much more expensive and time-comsuming.

2) Transferable-E2E-ABSA: we firstly explore an unsupervised domain adaptation (UDA) setting for cross-domain E2E-ABSA. Unlike the traditional UDA in classification problems, this task aims to leverage knowledge from a labeled source domain to improve the sequence learning in an unlabeled target domain.

Requirements

  • Python 2.7.12

  • Tensorflow-gpu 1.4.1

  • numpy 1.15.4

Environment

  • OS: CentOS Linux release 7.5.1804
  • GPU: NVIDIA TITAN Xp
  • CUDA: 8.0

Running

Download (Password: zlyc) the word embedding and then move it to the data directory. The embedding is pre-trained on Yelp and Electronics dataset.

AD-SAL (full Model):

selective adversairal learning on the low-level AD task.

python main.py --train --test -s rest -t service -model_name AD-SAL --selective

AD-AL (ablation Model):

pure adversairal learning without selectivity on the low-level AD task.

python main.py --train --test -s rest -t service -model_name AD-AL

Training over all transfer pairs:

./scripts/train_AD-AL.sh
./scripts/train_AD-SAL.sh

Citation

If the source code and data are useful for your research, please be kindly to give us stars and cite our paper as follows:

@article{li2019sal,
  title={Transferable End-to-End Aspect-based Sentiment Analysis with Selective Adversarial Learning},
  author={Li, Zheng and Li, Xin and Wei Ying and Bing Lidong and Zhang Yu and Yang, Qiang},
  conference={EMNLP},
  year={2019}
}
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].