Automatic Text Summarization is one of the most challenging topics in NLP. Abstractive Text Summarization is the task of creating summaries that contain new sentences than those already found in the original text, so the model needs to understand the language itself in order to create meaningful summaries without copying sentences from the original text.
In this project we use a seq2seq deep learning model with attention layer in the decoder to summarize long customer reviews on fine foods from Amazon into short concise reviews that delivers the main point of the review in the minimum number of words. More details about the implementation, trials, and results can be found in the report.