PWCN

PWCN - Proximity-Weighted Convolution Network

• Repo for SIGIR 2019 paper titled "Syntax-Aware Aspect-Level Sentiment Classification with Proximity-Weighted Convolution Network"
• Chen Zhang, Qiuchi Li and Dawei Song.

Requirements

• Python 3.6
• PyTorch 1.0.0
• SpaCy 2.0.18
• numpy 1.15.4

Usage

• Download pretrained GloVe embeddings with this link and extract glove.840B.300d.txt into glove/.
• Train with command, optional arguments could be found in train.py

python train.py --model_name pwcn_dep --dataset laptop

• Infer with infer.py

Model

We propose a proximity-weighted convolution network to offer an aspect-specific syntax-aware representation of contexts. In particular, two ways of determining proximity weight are explored, namely position proximity and dependency proximity. The representation is primarily abstracted by a bidirectional LSTM architecture and further enhanced by a proximity-weighted convolution.

An overview of our proposed model is given below

Docker Image

There is a Docker Image created to run this model easily. You can pull the image here

docker pull auliadil/pwcn

In order to run this docker image, just run

1. docker run -d --name pwcn auliadil/pwcn
2. Go to docker image terminal with docker exec -ti pwcn \bin\bash
3. Run conda activate myenv
4. Then, you can run the python train.py --model_name pwcn_dep --dataset laptop command

Citation

If you use the code in your paper, please kindly star this repo and cite our paper

@inproceedings{Zhang:2019:SAS:3331184.3331351,
 author = {Zhang, Chen and Li, Qiuchi and Song, Dawei},
 title = {Syntax-Aware Aspect-Level Sentiment Classification with Proximity-Weighted Convolution Network},
 booktitle = {Proceedings of the 42Nd International ACM SIGIR Conference on Research and Development in Information Retrieval},
 series = {SIGIR'19},
 year = {2019},
 isbn = {978-1-4503-6172-9},
 location = {Paris, France},
 pages = {1145--1148},
 numpages = {4},
 url = {http://doi.acm.org/10.1145/3331184.3331351},
 doi = {10.1145/3331184.3331351},
 acmid = {3331351},
 publisher = {ACM},
 address = {New York, NY, USA},
 keywords = {proximity-weighted convolution, sentiment classification, syntax-awareness},
}

Credits

• Code of this repo heavily relies on ABSA-PyTorch, in which I am one of the contributors.
• For any issues or suggestions about this work, don't hesitate to create an issue or directly contact me via gene_zhangchen@163.com !