Unified Feature and Instance Based Domain Adaptation for Aspect-Based Sentiment Analysis

This repository contains code for our EMNLP 2020 paper:

Unified Feature and Instance Based Domain Adaptation for Aspect-Based Sentiment Analysis

Datasets

The training data comes from four domains: Restaurant(R) 、 Laptop(L) 、 Service(S) 、 Devices(D).
For each domain transfer pairs, the unlabeled data come from a combination of training data from the two domains(ratio: 1:1).

The in-domain corpus(used for training BERT-E) come from yelp and amazon reviews.

Click here to get BERT-E (BERT-Extented) , and the extraction code is by0i. (Please specify the directory where BERT is stored in modelconfig.py.)

Dependencies

• Python 3. (test on 3.7.6)
• torch 1.5.1
• tqdm 4.42.1
• pytorch-transformers 1.2.0
• spacy 2.3.2

Usage

Pre-steps, run below code to get the data for auxiliary tasks:

bash ./scripts/create_aux_data.sh

To get the BERT-B-UDA results (based on bert-base-uncased) for all domain transfer pairs:

• Step 1, run:

bash ./scripts/run_base_feature_learning.sh

• Step 2, run:

bash ./scripts/run_base_uda.sh

To get the BERT-E-UDA results(based on bert-extented) for all domain transfer pairs:

• Step 1, run:

bash ./scripts/run_extented_feature_learning.sh

• Step 2, run:

bash ./scripts/run_extented_uda.sh

To get BERT-Base results, run:

bash ./scripts/run_base.sh

To get BERT-Extented results, run:

bash ./scripts/run_extented.sh