BERT for ABSA

Table of Contents

• About the Project
  • Built With
• Installation
• Usage
• Authors
• Acknowledgments

About The Project

Replication of the methodology proposed in the paper Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence.

In this work (Targeted) Aspect-Based Sentiment Analysis task is converted to a sentence-pair classification task and a pre-trained BERT model is fine-tuned on it.

More details about the project in the presentation.

Built With

• 🤗 Transformers
• PyTorch

Installation

To get a local copy up and running follow these simple steps:

1. Clone the repo

git clone https://github.com/LorenzoAgnolucci/BERT_for_ABSA.git

2. Run pip install -r requirements.txt in the root folder of the repo to install the requirements

Usage

1. Run generate_datasets.py to build the datasets corresponding to each model or simply use the ones provided in data/sentihood/ and data/semeval2014/

2. Use the forms in BERT_for_ABSA.ipynb to choose the desired dataset type and task both for BERT-single and BERT-pair. Then fine-tune the model and evaluate it runnning the corresponding cells

3. Run the subsequent cells in BERT_for_ABSA.ipynb to fine-tune and evalaute the model

Authors

• Lorenzo Agnolucci

Acknowledgments

Machine Learning © Course held by Professor Paolo Frasconi - Computer Engineering Master Degree @University of Florence