Skip to content

Automatic Design of Semantic Similarity Ensembles Using Grammatical Evolution

License

Notifications You must be signed in to change notification settings

jorge-martinez-gil/sesige

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Automatic Design of Semantic Similarity Ensembles Using Grammatical Evolution

arXiv

🌟 Introduction

Semantic similarity measures are essential in natural language processing, aiding a myriad of computer-related tasks. Our research introduces an innovative method for automatically designing semantic similarity ensembles using grammatical evolution, marking its first-time application in this domain.

📊 Features

  • Automatic Ensemble Design: Utilizes grammatical evolution for ensemble creation.
  • Dynamic Measure Selection: Aggregates various measures to form an optimized ensemble.
  • Benchmark Evaluations: Tested against top-tier ensembles on standard datasets.
  • Accuracy Improvements: Demonstrates notable enhancements in similarity assessments.

🛠️ Install

This research is heavily based in this work:

Fenton, M., McDermott, J., Fagan, D., Forstenlechner, S., Hemberg, E., and O'Neill, M. 
PonyGE2: Grammatical Evolution in Python. arXiv preprint, arXiv:1703.08535, 2017.

Therefore, for making it running it is necessary to install the PonyGE2 framework first:

  1. Install PonyGE2.
  2. Clone this repository.
  3. Overwrite the PonyGE2 files with the files from this repository.

📈 Datasets

We evaluated our approach on MC30 and GeReSiD50 datasets. For more details, refer to our paper.

⚙️ Usage

After installing the Pony2GE framework:

cd ./PonyGE2/src
python ponyge.py --parameters <your_parameter_file>

📚 Citation

If you use our work, please cite:

@article{martinez2003c,
  author       = {Jorge Martinez-Gil},
  title        = {Automatic Design of Semantic Similarity Ensembles Using Grammatical
                  Evolution},
  journal      = {CoRR},
  volume       = {abs/2307.00925},
  year         = {2023},
  url          = {https://doi.org/10.48550/arXiv.2307.00925},
  doi          = {10.48550/arXiv.2307.00925},
  eprinttype   = {arXiv},
  eprint       = {2307.00925}
}

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.