yukioichida/semantic-neural-network

Neural network model to measure semantic similarity between sentences.

Python

Siamese Neural Network

Neural network model for measure semantic similarity between two sentences using Siamese Architeture. DOI: https://doi.org/10.1109/IJCNN.2018.8489433

Requirements

Python 2.7
Keras 2.2.1
(see requirements.txt)

Files

run_model.py - Run the neural network training
modules/config - Project Configuration values (dataset files, hyperparameters)

Dataset used

SICK (Sentences involving compositional knowledge): https://github.com/brmson/dataset-sts/tree/master/data/sts/sick2014
Word2Vec - Negative Sampling: https://code.google.com/archive/p/word2vec/