This is the PyTorch implementation of work presented in 'Modelling Context with User Embeddings for Sarcasm Detection in Social Media' (https://arxiv.org/pdf/1607.00976.pdf). The neural network takes a tweet (content) and corresponding user embedding (context) as input, and classifies the tweets as sarcastic/non-sarcastic.

• python 2.7
• PyTorch 0.3.1
• python package gensim
• python package yandex.translate
• python package ipdb

1. Get pre-trained word embeddings (e.g. Skip-gram)

   • Install the bin file from this link
   • Unzip the .bin.gz fine and run the iPython notebook get_word2vec_embeddings.ipynb
   • Place the .txt file obtained in DATA/embeddings/ and change its name to words.txt

2. Get pre-trained user embeddings for the user. The embeddings we used can be found here. Place the embeddings in DATA/embeddings and name the file as usr2vec.txt

3. Execute iPython notebook get_data.ipynb. This utility code is used to download tweets corresponding to the tweet ids and then preprocess these tweet messages.

Run python train_CUE_CNN.py

Run python Headlines_RNN.py

The code generate a progress folder, that contains sub folder for every run. Inside every run folder following two file are generated -

1. logs.txt which contains loss and accuracy on train/test/validation set after every epoch
2. stats.jpg that plots
   • train/test/validation loss on a single plot
   • train/test/validation accuracy on a single plot

Util files, pre-trained user embeddings and raw tweet ids were obtained from Original CUE-CNN