Text Summarization with Different Encoders for Pointer Generator Network

The ever-growing increase of documents has compelled the need of text summarization. In the past, deep learning models have shown state-of-the-art results for text summarization. In this paper, a comparison is conducted between different encoders for pointer generator network. The two different encoders used for comparison are bi-directional GRU encoder and bi-directional LSTM encoder. The decoder used with both the encoders is unidirectional LSTM decoder. The results are evaluated using ROUGE value. The experiments show that the bi-directional LSTM gives better result in comparison to GRU encoder.